# Your First Day with SoloEnt
Source: https://docs.soloent.ai/en/beginner-guide

From download to installing your first plugin, in the order a new user actually does it

This guide follows the order a new user actually works through. Budget about 15 minutes.

## Step 1: Download

Go to [soloent.ai](https://soloent.ai/), click **Download** in the top-right corner, and pick the build for your machine.

<img alt="官网右上角「下载」入口" />

Windows users: check your chip architecture. Mac users: check whether you need the Apple Silicon or Intel build.

<img alt="下载页的 Windows / Mac 版本选择" />

## Step 2: Install

### Windows

You may see a security warning during installation. This is because SoloEnt is not distributed through the Microsoft Store — the installer itself is safe. Choose **Trust**, **Accept**, **Keep**,or **Run** to continue.

<Warning>
  Make sure no older SoloEnt window is still running before you install, or the installation will fail.
</Warning>

<img alt="Windows security warning during installation" />

### Mac

Double-click the installer and drag the app into your Applications folder.

<img alt="Mac 将应用拖入 Applications 文件夹" />

## Step 3: Set your language

The first time you open SoloEnt, the welcome page has a language switcher in the top-right corner.

<img alt="Language and Theme options in the top-right of the welcome page" />

Once you are in the app, the settings icon in the top-right of the Agent panel gives you two separate options:

* **Preferred Language** — the language the AI replies in
* **Interface Language** — the display language of the Agent panel itself

<img alt="Preferred Language and Interface Language options in the Agent settings panel" />

These two are independent, so an English interface with replies in another language is a valid combination.

## Step 4: Pick a color theme

If you write for long stretches, start with a theme that is easy on the eyes.

<img alt="Theme option in the top-right of the welcome page" />

In the theme picker, use the **arrow keys to preview themes live** and press **Enter** to confirm.

<img alt="Theme settings menu" />

## Step 5: Open or create a project

**In SoloEnt, one work equals one folder** stored locally on your machine. The first thing to do after launching the app is open a folder.

### Open an existing project

Select the project folder and click **Open**.

<Warning>
  Open the **folder**, not a single file. Opening a single file leaves the Agent without project context.
</Warning>

### Create a new project

In the file picker, navigate to where you want the work to live, click **New Folder** in the bottom-left, then open it.

<img alt="文件选择弹窗中的「新建文件夹」按钮" />

Once it is open, the Agent will look for and create files in that folder as your writing requires — world, characters, outline, chapters, and so on.

## Step 6: Learn the interface

The interface has three panels, left to right. Every divider is draggable. The four icons in the top-right switch between layout presets.

<img alt="Three-panel interface: file manager, editor, and Agent panel" />

| Panel               | What it does                                                                                                                                                                  |
| ------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Left — file manager | Manages your work: world, characters, outline, chapters, plus anything else you or the AI generate                                                                            |
| Center — editor     | Read and edit documents, primarily Markdown. Supports split and multi-pane views, and formatted preview                                                                       |
| Right — Agent       | Talk to the Agent to run writing tasks, switch between Plan and Act modes, drag files in from the left, quote selected text from the editor, and call subagents or web search |

<Tip>
  Two habits worth forming early: **drag files into the chat box**, and **select text in the editor and quote it into the chat**. Both are cheaper in tokens and more precise than copy-paste.
</Tip>

## Step 7: Sign in

You need to be signed in to use the official models and your free credits.

<Steps>
  <Step title="Click your avatar">
    Click the avatar in the top-right of the Agent panel, then sign in.

    <img alt="Agent 面板右上角的账号头像入口" />
  </Step>

  <Step title="Sign in through the browser">
    The login page opens automatically in your browser. Sign in with Google or Microsoft.
  </Step>

  <Step title="Return to the app">
    After signing in, the page asks whether to open SoloEnt. Click **Open**.

    <img alt="Authorization success page with the browser prompt to open SoloEnt, and the fallback Click here to open SoloEnt button" />
  </Step>

  <Step title="Approve the request">
    Back in the app, an authorization prompt appears. Click **Open** to finish.

    <img alt="客户端内的授权请求弹窗" />
  </Step>

  <Step title="Set an email and password (optional)">
    After your first successful third-party sign-in, you can set an email and password in the dashboard and use those from then on.
  </Step>
</Steps>

## Step 8: Configure your API

This is where new users most often get stuck, but **most people do not need to configure anything** — sign in and the free official models are ready to use.

### Two ways in

* **Quick** — click the model name at the bottom of the Agent chat box to open the model settings directly
* **Full** — click the settings icon in the Agent panel and go to API Configuration

The configuration page has two columns: **Official** (no setup, just pick a model) and **BYOK** (bring your own API key or a local model).

<img alt="Quick API configuration from the model name in the chat box" />

<img alt="API configuration via the settings panel" />

### Option A: Use the official API (recommended for beginners)

**No configuration needed. Sign in and go.**

* Free models are provided, and every registered user gets **US\$0.5** in trial credit
* More free credit is available through platform activities
* The free model list and any discounts change with ongoing promotions
* Lite, Pro, and Max subscribers get every model in the VIP list, always the latest official releases. Membership switches you to the VIP list automatically
* Official API pricing matches each model vendor's own pricing, with **no surcharge**

You can search the model list by name or vendor. Each model shows its context length and input/output pricing.

<img alt="API configuration panel showing the official provider and the Free tab with available models" />

### Option B: Bring your own third-party API key (BYOK)

First, one thing to be clear about: **an API key is not the same as a subscription on the model vendor's website.**

Many people assume a US\$20 monthly plan on a vendor's site includes API access. It does not — the two are billed separately. An API key is:

* Your credential for calling the model
* The thing your usage is billed against
* A secret worth protecting. It looks like `sk-abc123xyz456...` — never share it, never commit it to a public repository

To configure:

<Steps>
  <Step title="Select BYOK">Choose BYOK on the configuration page.</Step>
  <Step title="Pick your API provider">Select your provider from the dropdown.</Step>
  <Step title="Enter your credentials">Fill in the values your provider gave you.</Step>
</Steps>

<Warning>
  Only keys bought directly from a model vendor, and a small number of aggregators, can be configured by picking a provider. **Most third-party providers require the OpenAI Compatible option, which needs Base URL, API Key, and Model ID — all three.** Confirm those three values with your provider before you buy.
</Warning>

<img alt="BYOK 下选择 OpenAI Compatible，需填 Base URL、API Key、Model ID" />

**Buy only from official sources, and be careful with payment details.**

### Option C: Reuse an existing Codex or Claude Code subscription

If you already pay for either one, you can spend that subscription inside SoloEnt and still get SoloEnt's writing Skills.

**Codex** — choose BYOK, set the provider to **OpenAI Codex**, click Sign in to OpenAI Codex, then pick a Codex model from the model list.

<img alt="BYOK configuration with API Provider set to OpenAI Codex, showing the Sign in to OpenAI Codex button" />

**Claude Code** — install the Claude Code CLI locally and sign it into your subscription first. Then choose BYOK, set the provider to **Claude Code**, and pick the Claude model you want. No other setup is needed.

<img alt="BYOK configuration with API Provider set to Claude Code, showing the CLI path and model selection" />

### Advanced: different models for Plan and Act

Official models can be assigned separately to planning and execution:

* Planning rewards structure and good clarifying questions — Sonnet and GLM do well here
* Execution rewards speed, concision, or lower cost — Gemini and Doubao do well here

<img alt="API Configuration panel with Plan Mode and Act Mode tabs and the option to use different models for each" />

<Info>
  This is optional. It suits users with a strong feel for individual models — if you are new, leave the defaults alone.
</Info>

## Step 9: Install your first plugin

Plugins are how SoloEnt's writing capabilities are distributed. A plugin may contain a Rule, a Workflow, a Skill, or any combination, aimed at one specific writing scenario. **You never have to create configuration files by hand — install it and follow its instructions.**

You can browse everything published at [soloent.ai/en/plugins](https://soloent.ai/en/plugins).

### Open a project first (this matters)

<Warning>
  Plugins install into **the currently active project**, not a global directory. Open the writing project you want to use it in (Step 5) before installing. Installing with no project open puts the files in the wrong place.
</Warning>

### Open the Marketplace

Click the Marketplace icon (the cube) at the top of the Agent panel.

<img alt="插件市场页面，顶部立方体图标为入口" />

Two ways to browse:

* Type a keyword in the search bar
* Use the language filter in the top-right (English, 中文, 日本語, 한국어, Español)

### Read the detail page

Open a plugin to see its detail page. The things worth reading:

* **Name and summary** — one or two sentences on the problem it solves
* **Install count** — a rough signal of how well it works
* **Usage** — how to invoke or trigger it
* **Contents** — which of Rule / Workflow / Skill it includes
* **Core rules** — the logic behind it, at a glance
* **Version** — the currently available version

### Install and verify

Click **Install** on the detail page. Installed plugins appear under the **Installed** tab.

Depending on what the plugin contains, verify the files landed:

| Contents | Where to check             |
| -------- | -------------------------- |
| Rule     | Agent panel → Rules tab    |
| Workflow | Agent panel → Workflow tab |
| Skill    | Agent panel → Skills tab   |

A combined plugin will show up in more than one place.

### Start using it

The three content types are triggered differently:

* **Rule** — active as soon as it is installed, nothing to call
* **Workflow** — triggered by its command
* **Skill** — triggered automatically when the situation calls for it

Every plugin's detail page documents its trigger, its prerequisites (for example, having a chapter file open), and what it produces. **Read that page fully before you start.**

## After that: three things worth doing immediately

### 1. Understand Plan mode vs Act mode

Switch between them in the bottom-left of the Agent chat box.

<img alt="Plan and Act mode selector in the Agent chat box" />

* **Plan** — produces a proposal without touching files. Use it when starting a book, when you are stuck, or when restructuring. It analyzes, breaks the work into steps, and asks for confirmation, ending with a plan you have approved.
* **Act** — does the work: creates files, edits text, runs commands. Day-to-day writing stays in Act.

<Tip>
  If the Agent only gives opinions and never generates files, check whether you are still in Plan mode.
</Tip>

A good rhythm: start a new project in Plan, switch to Act once the plan is agreed, and go back to Plan when you are stuck or making a large change.

### 2. Initialize SOLOENT.md

Type `/init` in the chat and SoloEnt generates a `SOLOENT.md` for the current project. It is the project's control panel and the AI's long-term memory, covering eight sections: project DNA, world system, characters, plot and structure, style guide, constraints and taboos, live writing state, and roadmap.

It exists to solve consistency over long works. **Short projects and non-fiction do not need it.**

As you write, the Agent maintains the live writing state and roadmap sections on its own. The two you should edit by hand are the style guide and the constraints.

### 3. Where to get help

<CardGroup>
  <Card title="FAQ" icon="circle-question" href="/en/faq">
    Answers to the most common questions
  </Card>

  <Card title="Troubleshooting" icon="wrench" href="/en/troubleshooting/login">
    Login, API, file format, and environment issues
  </Card>

  <Card title="Email support" icon="envelope" href="mailto:support@soloent.cn">
    [support@soloent.cn](mailto:support@soloent.cn)
  </Card>

  <Card title="Subscription and payment" icon="credit-card" href="/en/subscription/choose-your-plan">
    Plans, checkout, and managing your subscription
  </Card>
</CardGroup>


# FAQ
Source: https://docs.soloent.ai/en/faq

Quick answers to common SoloEnt questions

Below are illustrated answers to common questions — scroll to find yours. If your question isn't listed, check the corresponding feature page in this handbook for details.

<Note>
  **Still stuck?** Email [team@soloent.ai](mailto:team@soloent.ai) with the following:

  1. Your login account (email address)
  2. Your client version number
  3. Detailed steps to reproduce
  4. Relevant screenshots
</Note>

## The login redirect prompt does not appear

After authorization succeeds, the browser usually asks whether to open SoloEnt, and there is a backup button at the bottom of the page:

* In the popup, choose **Open SoloEnt** to allow the desktop app to open
* If there's no popup, click **Click here to open SoloEnt** on the page to trigger it manually

<img alt="Authorization success page: the browser asks whether to open SoloEnt, with the Click here to open SoloEnt backup button on the page" />

If there's still no response, the browser is most likely blocking the popup. In Chrome, for example: **Settings → Privacy and security → Site settings** (which controls a site's use of location, camera, **pop-ups**, and more), then allow pop-ups and redirects for `account.soloent.ai`.

<img alt="Chrome settings: Privacy and security and the Site settings entry" />

In **Site settings**, turn on **Pop-ups and redirects**, set SoloEnt's sites to allowed, and retry the login.

<img alt="The Pop-ups and redirects option in Site settings, set to allow sites to send pop-ups and use redirects" />

## Can't find the recharge/upgrade entry

There are two ways to reach the subscription console:

* **Option 1**: Click your account avatar to open the menu and select **Dashboard**, which opens [account.soloent.ai/en/dashboard](https://account.soloent.ai/en/dashboard/)

<img alt="Reaching the console via Dashboard from the account avatar menu" />

* **Option 2** (from v0.3.5): Click the Agent **Settings** in the top-right → **API settings**; there's an upgrade-to-VIP prompt and entry above the free model list.

<img alt="Opening API Configuration in Settings, with the Upgrade subscription button below the free model list" />

Once in the web console (Dashboard), switch to the **Subscription Plan** tab and choose your plan (Lite, Pro, or Max) and billing cycle (monthly or yearly).

<img alt="Choosing a plan and billing cycle under the Subscription Plan tab in the console" />

## Can I get a refund? What if I run out of credits?

In the console, the button for your currently active plan shows **Cancel subscription**; click and confirm to cancel.

* After canceling, the current period is **not refunded**; your benefits continue until the period ends and it won't auto-renew.
* If your subscription credits run out, you can upgrade to a higher tier, or **buy add-on credits separately** (note: without an active membership you cannot top up credits directly).

<img alt="The console Subscription Plan page: the Cancel subscription and Buy add-on quota entries under the Lite plan" />

<Warning>
  Add-on credits apply only within the current billing cycle. When your subscription renews next month or you make a subscription change (upgrade, downgrade, switch period), your quota resets to the plan's base amount, and **top-up credits do not carry over**.
</Warning>

For more on managing your subscription, see [Manage subscription](/en/subscription/manage-subscription) and [Make the payment](/en/subscription/make-the-payment).

## How do I configure a third-party API key?

SoloEnt supports BYOK for free. Just select **BYOK** on the **API settings** page, then choose your API provider from the dropdown.

<img alt="Selecting BYOK in API settings and choosing a provider from the API Provider dropdown" />

Most third-party providers' keys need to be configured via **OpenAI Compatible**.

You must fill in all three fields:

* **Base URL**
* **API Key**
* **Model ID**

<img alt="Under BYOK, API Provider set to OpenAI Compatible, with Base URL, API Key, and Model ID to fill in" />

If it doesn't work after configuring, first check that all three are entered correctly. Confirm with your provider before purchasing, as some third-party services can be unstable.

Most official providers' APIs can also be configured via **OpenAI Compatible**, as long as the Base URL you enter is correct.

<Warning>
  If you use a third-party API, we **cannot help resolve API call errors**. When you hit errors, try switching to a more stable API.
</Warning>

For details, see [API Configuration](/en/quickstart/api-configuration).

## The Agent panel language isn't what I want

SoloEnt is the international edition. After you change the interface language on the welcome page, only some prompts in the main window change. The right-side Agent's language must be set separately.

<img alt="The Language and Theme options in the top-right of the welcome/main interface" />

Click the settings icon in the top-right of the Agent panel and configure, under General Settings:

* **Preferred Language**: the language the AI replies in — the language you receive responses in when chatting with the Agent
* **Interface Language**: the display language of the Agent panel itself (menus, buttons, etc.); supports 简体中文, English, 日本語, 한국어, Español

<img alt="The Preferred Language and Interface Language options under the Agent panel's General Settings" />

## The Agent won't create/save files, or always asks me to save

Check the following two settings in order.

**1. Auto-approve**\
In the **Auto-approve** settings at the bottom-right of the Agent window, allow **auto-edit** so the Agent can write to project files without your approval.

If it's off, every file write needs your confirmation. See [Auto-approve](/en/quickstart/auto-approve).

<img alt="Permission toggles for reading and editing project files in the Auto-approve settings" />

**2. Background Edit**

In **Agent Settings → Tool settings** (the second item on the left), the **Background Edit** toggle under the **Editor** block controls whether files can be edited in the background **without taking editor focus**.

If you always want to review before revising, keep this off. Otherwise the Agent may modify files without your noticing.

<img alt="The Background Edit toggle in the Editor block of Feature Settings" />

## Prompt to install Git (for checkpoints)

The first time Windows users open SoloEnt, they may see the prompt shown below. It's because Git isn't installed, but it **doesn't affect normal use**.

Git is used to save state checkpoints during a conversation so you can roll back to a checkpoint and start over. macOS usually ships with Git; Windows usually needs a manual install. If you don't need checkpoint recovery, you can ignore this prompt for now; when you do need it, download and install from [git-scm.com](https://git-scm.com).

<img alt="The Git and checkpoints prompt banner in the SoloEnt panel (Git must be installed to use checkpoints)" />

## What is the context window? When should I open a new one?

Every session window has a context-length limit.

Above the chat box you can see the current window's session ceiling and how much context is used — for example the **30.2k / 300.0k** below: the left number is the tokens used, and the right number is the current session's context ceiling.

<img alt="The context progress bar above the SoloEnt chat box, showing 30.2k used and a 300.0k ceiling" />

When a lot of context is used, the model's memory of earlier content gets fuzzy and it tends to drift on long tasks. And **bigger isn't always better** — an oversized window scatters the model's attention and pushes up cost, so keep the session ceiling around **200K–300K**.

Consider opening a new window when:

* The current task is done and you're starting a separate one
* A single window has reached about **100–150K** (continuing past 200K may cause a cost surge)
* You're switching to a new direction unrelated to the current topic

<Tip>
  As a rule of thumb, once a single window passes 100k, it's worth opening a new one.
</Tip>

Before opening a new window, you can use `/compact` to condense the current session into a short summary, or have the Agent summarize the key points into `SOLOENT.md`, so the new window quickly restores context instead of piling up tokens from scratch.

For details, see [Using the Context Window Wisely](/en/tips/context-window) and the [SOLOENT.md Guide](/en/tips/SoloEnt).

## Why doesn't my project generate SOLOENT.md automatically anymore?

As of v0.2.0, `SOLOENT.md` is **no longer generated automatically**. You can create it manually at any stage of a project with the `/init` command, or skip it entirely for short pieces or when you don't need context tracked across sessions.

## Why do files sometimes fail midway or get regenerated repeatedly?

Common causes include:

* **The write was interrupted by hand**: while the Agent is editing a file, editing it yourself, closing the Agent's edit box, or making changes while the model is thinking can all cause duplication or truncation
* **BYOK model output is truncated**: a third-party model's max output length may be capped, so the truncated result makes the file-writing tool call fail
* **The client is outdated**: older clients support newer models poorly — check your client version and keep it updated
* **The single-window context is too long**: when context is too long, the model's tool-call output can be malformed. Tidy up your `SOLOENT.md` memory file and continue in a new session, rather than staying in one session indefinitely


# SoloEnt Handbook
Source: https://docs.soloent.ai/en/introduction

The intelligent workspace for AI-native creators

## What is SoloEnt?

SoloEnt is an AI-powered desktop writing tool designed for creative writers.

We champion the philosophy of **vibe writing**, empowering AI-assisted creation through practical tools and shared knowledge.

<Note>
  AI writing isn't a single path, nor is it limited to one tool. The only constant is the **personal growth and continuous evolution** of every creator in the age of AI.
</Note>

## QuickStart

<CardGroup>
  <Card title="Quick Start" icon="rocket" href="./quickstart/quickstart">
    Get up and running in 5 minutes
  </Card>

  <Card title="Create Your First Project" icon="folder-plus" href="./tips/create-project">
    Set up new writing projects
  </Card>

  <Card title="Marketplace" icon="store" href="./tips/marketplace">
    Install official writing skill packs in one click
  </Card>

  <Card title="Join Discord" icon="discord" href="https://discord.com/invite/NUh3skEFuY">
    Connect with the community and grow together
  </Card>
</CardGroup>

## Key concepts

<AccordionGroup>
  <Accordion title="Plan Mode vs Act Mode">
    * **Plan Mode**: Think through complex tasks and create strategies before taking action
    * **Act Mode**: Execute tasks immediately for straightforward operations
    * Learn more in the [Plan Mode guide](./tips/plan-mode)
  </Accordion>

  <Accordion title="SOLOENT.md - Your project constitution">
    A comprehensive document that defines your story's DNA, characters, world, and writing guidelines. The Agent uses this as reference for all writing tasks.

    * Learn more in the [SOLOENT.md guide](./tips/SoloEnt)
  </Accordion>

  <Accordion title="Marketplace">
    Official writing skill packs including Rules, Workflows, Skills, and combinations — install in one click to extend your Agent's capabilities.

    * Learn more in the [Marketplace guide](./tips/marketplace)
  </Accordion>

  <Accordion title="Using the context window">
    Every session has a context ceiling, and bigger isn't always better. Managing session length and opening new windows at the right time keeps long-form writing stable and saves tokens.

    * Learn more in the [Using the Context Window Wisely guide](./tips/context-window)
  </Accordion>
</AccordionGroup>

## Quick navigation

<Tabs>
  <Tab title="I'm new">
    Start here if you're using SoloEnt for the first time:

    1. [Quick Start](./quickstart/quickstart) - Setup in 5 minutes
    2. [Create Your First Project](./tips/create-project) - Start writing
    3. [Plan Mode](./tips/plan-mode) - Learn strategic planning
    4. [SOLOENT.md Guide](./tips/SoloEnt) - Understand the methodology
  </Tab>

  <Tab title="I need help">
    Having issues? Find solutions:

    * [Troubleshooting Index](./troubleshooting) - Common problems and fixes
    * [API Configuration Issues](./troubleshooting/api-configuration) - API setup problems
    * [Environment Issues](./troubleshooting/environment) - System and shell problems
    * Email: [team@soloent.ai](mailto:team@soloent.ai)
  </Tab>

  <Tab title="I want free options">
    Writing on a budget:

    * [Free API Keys](./resources/free-api-keys) - Free tier access
    * [Local LLMs](./resources/local-llms) - Run models locally
    * [API Directory](./resources/api-directory) - Compare providers
  </Tab>
</Tabs>

## Community & support

<CardGroup>
  <Card title="Discord Community" icon="discord" href="https://discord.com/invite/NUh3skEFuY">
    Join our writers community
  </Card>

  <Card title="X (Twitter)" icon="x-twitter" href="https://x.com/soloent_ai">
    Follow for updates
  </Card>

  <Card title="Facebook" icon="facebook" href="https://www.facebook.com/profile.php?id=61588035750230">
    Connect on Facebook
  </Card>
</CardGroup>

<Tip>
  **Pro tip**: Use the search bar at the top to quickly find any topic in this handbook.
</Tip>


# API Configuration
Source: https://docs.soloent.ai/en/quickstart/api-configuration

Set up AI model API keys in SoloEnt

Configure API keys to enable SoloEnt's AI writing features.

## Entry points

### Quick config from chat

<Steps>
  <Step title="Click the model name at the bottom of the chat">
    Click the model name at the bottom of the Agent chat to open the model settings panel.
  </Step>
</Steps>

<img alt="Quick API configuration from chat" />

### Via settings panel

<Steps>
  <Step title="Open Agent settings">
    Click the settings icon in the Agent panel.
  </Step>

  <Step title="Go to model settings">
    Choose whether to use official models directly or configure your own model (BYOK).
  </Step>
</Steps>

<img alt="API configuration via settings panel" />

## Official API

SoloEnt provides an official API. Once logged in, you can use it directly without setting up any third-party keys.

### Free models

Official free models are available immediately after login. The available free models and discounts are updated regularly with ongoing campaigns.

<Note>
  Every registered user gets a trial credit of **\$0.5**; join [platform activities](https://soloent.ai/en/activities) to earn more free credit.
</Note>

<img alt="Settings panel showing SoloEnt AI as the API provider with Free tab listing available free models" />

<Warning>
  Free models are for trying out the tool. Response times may be slower or less consistent during periods of high demand.
</Warning>

### Subscription models (VIP)

Lite, Pro, and Max subscribers can use all models in the VIP list — covering the latest models from top providers including Anthropic, OpenAI, Google, and DeepSeek — without configuring any third-party API key.

<Note>
  SoloEnt API pricing matches the official pricing of each model provider, with no additional fees, and different models offer different discounts from time to time.
</Note>

See all official models and the latest discounts at [soloent.ai/en/models](https://soloent.ai/en/models).

***

## Third-party API key configuration

An API key is your authentication credential for accessing AI models — typically a random string:

```
sk-abc123xyz456def789ghi012jkl345mno678pqr901stu234
```

<Warning>
  **Important:** API access is billed separately from monthly subscriptions. A \$20 subscription does not include API usage. For example, a Gemini membership you subscribe to cannot be used as an API here.
</Warning>

### Where to get API keys

#### Popular providers

| Provider        | Latest Models                      | Best For                                            |
| --------------- | ---------------------------------- | --------------------------------------------------- |
| **OpenAI**      | GPT-5.6                            | Writing research, general writing                   |
| **Anthropic**   | Claude Sonnet-5 / Opus-5 / Fable-5 | Story planning, strategic thinking, prose writing   |
| **Google**      | Gemini-3.1-pro                     | Fast, cost-effective, prose writing, multimodal     |
| **xAI**         | Grok-4.5                           | Real-time info, flexible content policy, multimodal |
| **DeepSeek**    | DeepSeek-V4-pro                    | Budget-friendly, flexible content policy            |
| **Zhipu AI**    | GLM-5.2                            | Story planning, Chinese content                     |
| **Moonshot AI** | Kimi-K3                            | Long context, Chinese content                       |
| **OpenRouter**  | Multiple models                    | Unified access, model comparison                    |

<Tip>
  See the [complete API directory](../resources/api-directory) for detailed provider information and registration steps.
</Tip>

<Card title="Get free API keys" icon="gift" href="../resources/free-api-keys">
  Explore free tier options and community resources
</Card>

### Filling in the API details you purchased

#### Standard configuration (official providers)

For OpenAI, Anthropic, Google, and select aggregators — find the corresponding provider in the API Provider dropdown and enter only the API Key:

```plaintext theme={null}
API Key: sk-abc123xyz456...
```

#### OpenAI-compatible configuration (third-party)

Select **OpenAI Compatible** from the API Provider dropdown to configure a third-party provider.

<img alt="API Configuration panel with API Provider dropdown expanded, OpenAI Compatible option highlighted" />

Most third-party providers require all three fields:

<ParamField type="string">
  API endpoint URL (e.g., `https://api.provider.com/v1`)
</ParamField>

<ParamField type="string">
  Your API key
</ParamField>

<ParamField type="string">
  Model identifier (e.g., `gpt-4-turbo`)
</ParamField>

<Warning>
  All three fields are required. Confirm with your provider before purchasing.
</Warning>

#### Local LLMs

Run AI models on your own computer for privacy and offline access.

| Solution      | Models                    | Best For                      |
| ------------- | ------------------------- | ----------------------------- |
| **Ollama**    | Llama 3.1, Qwen, DeepSeek | Easy setup, macOS/Linux       |
| **LM Studio** | Multiple open models      | GUI interface, cross-platform |
| **vLLM**      | Custom deployments        | High performance, production  |

<Card title="Local model setup guide" icon="laptop" href="../resources/local-llms">
  Complete guide to running models locally
</Card>

## Using a Codex / Claude Code subscription in SoloEnt

If you already have an OpenAI Codex or Claude Code subscription, you can reuse that subscription's quota directly in SoloEnt while enjoying SoloEnt's professional writing Skills — no need to buy a separate API key.

#### Using a Codex subscription

Select **BYOK**, choose **OpenAI Codex** in the API Provider list, and click to sign in to OpenAI Codex. Once signed in, you can select the models Codex provides (such as GPT-5.x) from the model list. No extra API key is needed when using a ChatGPT Plus or Pro subscription.

<img alt="BYOK configuration with API Provider set to OpenAI Codex, showing the Sign in to OpenAI Codex button" />

#### Using a Claude Code subscription

First install the **Claude Code CLI** on your computer and bind your Claude Code subscription account, making sure Claude Code runs correctly locally. Then select **BYOK**, choose **Claude Code** from the list, and — with no other setup — pick the Claude model you want directly from the model list.

<img alt="BYOK configuration with API Provider set to Claude Code, showing the Claude Code CLI Path and model selection" />

## Common error codes

| Cause                | Error Code | Error Message                                          |
| -------------------- | :--------: | ------------------------------------------------------ |
| Missing model        |    `400`   | `Missing model`                                        |
| Budget exceeded      |    `403`   | `Budget exceeded`                                      |
| Free quota exhausted |    `403`   | `Daily free quota used up, please try again tomorrow.` |
| Model not found      |    `404`   | `Model not found`                                      |

<Note>
  Other errors caused by instability in third-party models cannot be accurately diagnosed by SoloEnt at this time.
</Note>

## Other FAQ

<AccordionGroup>
  <Accordion title="Invalid API key">
    * Verify you copied the entire key
    * Check if the key has expired
    * Regenerate from the provider dashboard
  </Accordion>

  <Accordion title="Connection fails">
    * Check your internet connection
    * Verify the Base URL (for OpenAI Compatible configuration)
    * Try without a VPN
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup>
  <Card title="Language settings" icon="language" href="./language-settings">
    Set your preferred language
  </Card>

  <Card title="Start writing" icon="pen-to-square" href="../tips/create-project">
    Create your first project
  </Card>

  <Card title="Local models" icon="laptop" href="../resources/local-llms">
    Run AI models on your computer
  </Card>

  <Card title="Free API keys" icon="gift" href="../resources/free-api-keys">
    Get free access to AI models
  </Card>
</CardGroup>


# Auto-approve settings
Source: https://docs.soloent.ai/en/quickstart/auto-approve

Control which actions SoloEnt can take without asking for your approval each time

By default, SoloEnt asks you to approve every action it takes—reading files, saving content, browsing the web. Auto-approve lets you skip those confirmation prompts for actions you trust.

<img alt="自动审批设置面板，显示读取项目文件、编辑项目文件、执行安全命令、执行所有命令、使用浏览器、使用 MCP 服务器和启用通知等复选框" />

<Note>
  Note that if you do not enable the **Edit project files** permission, the Agent will ask for your confirmation **every time** it creates or modifies a file.
</Note>

<img alt="Auto-approve settings panel showing checkboxes for Read project files, Edit project files, Execute safe commands, Execute all commands, Use the browser, Use MCP servers, and Enable notifications" />

## Recommended setup for writers

Most writers do well with this baseline:

<CardGroup>
  <Card title="Always enable" icon="circle-check">
    **Read project files** — lets SoloEnt read your chapters, outlines, and character sheets without asking each time
  </Card>

  <Card title="Enable if you want SoloEnt to save" icon="floppy-disk">
    **Edit project files** — lets SoloEnt write content directly to your files, so nothing gets lost
  </Card>

  <Card title="Enable for workflows" icon="terminal">
    **Execute safe commands** — workflows rely on commands to run and save their reports automatically
  </Card>

  <Card title="Enable for research" icon="globe">
    **Use the browser** — lets SoloEnt look up facts, names, or references while you write
  </Card>
</CardGroup>

<Note>
  **Execute all commands** and **Use MCP servers** are not needed for writing work — leave them off.
</Note>

## What each permission does

| Permission                | What it means for writers                                                                             |
| ------------------------- | ----------------------------------------------------------------------------------------------------- |
| **Read project files**    | SoloEnt can read your chapters, outlines, and character sheets as context                             |
| Read all files            | SoloEnt can read files outside your project folder                                                    |
| **Edit project files**    | SoloEnt saves generated content directly to your files                                                |
| Edit all files            | SoloEnt can edit files outside your project folder                                                    |
| **Execute safe commands** | Powers workflows — reads project files, runs the workflow, and saves reports to the required location |
| Execute all commands      | Runs any system command — not needed for writing                                                      |
| **Use the browser**       | SoloEnt can look up information online while helping you write                                        |
| Use MCP servers           | Advanced tool integrations — not needed for most writers                                              |
| Enable notifications      | Sends an alert when a long-running task finishes                                                      |

<Note>
  "Read all files" and "Edit all files" only extend access outside your project folder. If the base toggle is off, these sub-options have no effect.
</Note>

## Common writing scenarios

<Accordion title="I want to use workflows">
  Enable **Execute safe commands**. SoloEnt's built-in workflows (like `/review` for chapter quality checks and `/market` for publishing strategy) need this permission to read your project files, run the workflow, and save reports to the `notes/` folder. Without it, every step pauses for manual confirmation.
</Accordion>

<Accordion title="I'm worried about losing content after a long generation">
  Enable **Edit project files**. SoloEnt saves content directly to your files as it generates—you won't lose work if you close the window or something interrupts the session.

  <Warning>
    With edit access enabled, SoloEnt can overwrite existing content. Review changes carefully, and use the chat history to recover anything unexpected.
  </Warning>
</Accordion>

<Accordion title="I want SoloEnt to research while I write">
  Enable **Use the browser**. SoloEnt can look up historical facts, place names, terminology, or any reference material without you having to switch tabs.
</Accordion>

<Accordion title="I just want to chat and brainstorm without any file access">
  Leave everything off. SoloEnt can still talk through ideas, suggest plot directions, and help you think—it just won't read or touch your files unless you paste content directly into the chat.
</Accordion>

<Accordion title="I want SoloEnt to work through a long task while I step away">
  Enable **Enable notifications**. You'll get an alert when the task finishes, so you don't have to keep checking back.
</Accordion>

## Next steps

<CardGroup>
  <Card title="Interface overview" icon="grid-2" href="./interface">
    Learn the three-panel workspace
  </Card>

  <Card title="API configuration" icon="key" href="./api-configuration">
    Set up your API keys
  </Card>
</CardGroup>


# Interface Overview
Source: https://docs.soloent.ai/en/quickstart/interface

Understanding SoloEnt's three-panel workspace

SoloEnt uses a three-panel layout designed for efficient writing with AI assistance.

## Three-panel layout

<CardGroup>
  <Card title="Left: Files" icon="folder-tree">
    Project structure and navigation
  </Card>

  <Card title="Center: Editor" icon="pen-to-square">
    Writing and editing workspace
  </Card>

  <Card title="Right: Agent" icon="robot">
    AI assistant and chat
  </Card>
</CardGroup>

<img alt="SoloEnt three-panel interface showing file manager on left, editor in center with chapter content, and Agent chat on right" />

<Tip>
  Drag panel boundaries to adjust each section. Changes save automatically for your next session.
</Tip>

## File manager (Left panel)

Organize your writing project with standard folders:

* **World**: World-building, settings, background information
* **Characters**: Character profiles and development
* **Outline**: Story structure and plot points
* **Chapters**: Your actual written content
* **Custom**: Any additional documents you create

<Note>
  Keep files organized by category. This helps both you and the AI understand your project better.
</Note>

### File operations

* **Create**: Right-click folder → **New File**
* **Rename**: Right-click file → **Rename**
* **Move**: Drag and drop between folders
* **Delete**: Right-click file → **Delete**

## Editor (Center panel)

Your main writing workspace with essential features:

<CardGroup>
  <Card title="Markdown support" icon="markdown">
    Write with standard Markdown formatting
  </Card>

  <Card title="Split view" icon="table-columns">
    Open multiple files side-by-side
  </Card>

  <Card title="Live preview" icon="eye">
    See formatted content in real-time
  </Card>

  <Card title="Auto-save" icon="floppy-disk">
    Changes save automatically
  </Card>
</CardGroup>

<Info>
  Learn more in the [Editor guide](../tips/Editor).
</Info>

## Agent (Right panel)

Your AI writing assistant with two modes:

<Tabs>
  <Tab title="Plan Mode">
    **For strategic thinking:**

    * Brainstorm ideas
    * Develop outlines
    * Research topics
    * Plan story arcs
  </Tab>

  <Tab title="Act Mode">
    **For content creation:**

    * Generate text
    * Rewrite passages
    * Expand sections
    * Edit and refine
  </Tab>
</Tabs>

### Using the Agent

<Steps>
  <Step title="Reference files">
    Drag a file from the project on the left into the chat to provide context.
  </Step>

  <Step title="Chat naturally">
    Ask questions or give instructions in plain language.
  </Step>

  <Step title="Switch modes">
    Toggle between Plan and Act modes based on your task.
  </Step>
</Steps>

<Info>
  Learn more in the [Plan Mode guide](../tips/plan-mode).
</Info>

## Layout presets

Click the layout icons in the top-right corner to quickly adjust your workspace:

| Preset           | Use Case                   |
| ---------------- | -------------------------- |
| **Editor focus** | Maximize writing space     |
| **Balanced**     | Equal space for all panels |
| **Agent focus**  | Expand AI chat area        |
| **Custom**       | Drag boundaries freely     |

<Tip>
  Switch layouts based on your current task. Use Agent focus for planning, Editor focus for writing.
</Tip>

## Next steps

<CardGroup>
  <Card title="Log in" icon="right-to-bracket" href="./login">
    Sign in to get started
  </Card>

  <Card title="Language settings" icon="language" href="./language-settings">
    Set your preferred language
  </Card>

  <Card title="Theme settings" icon="palette" href="./theme-settings">
    Customize your color theme
  </Card>

  <Card title="API configuration" icon="key" href="./api-configuration">
    Configure your API keys
  </Card>
</CardGroup>


# Language Settings
Source: https://docs.soloent.ai/en/quickstart/language-settings

Configure interface and Agent language

Set your preferred language for the SoloEnt interface and AI Agent.

## Two language settings

<CardGroup>
  <Card title="Interface language" icon="desktop">
    Menus, buttons, and UI elements
  </Card>

  <Card title="Agent language" icon="robot">
    AI conversation language
  </Card>
</CardGroup>

## Setting interface language

### On first launch

The welcome screen lets you choose your interface language.

<img alt="SoloEnt welcome page with Choose Theme and Change Language options" />

## Setting Agent language

The Agent language must be set independently.

Click the **settings icon** in the top-right corner of the Agent panel. Under General Settings, you can configure the following two options:

* **Preferred Language**: Sets the language AI uses when replying — this is the language you receive responses in during conversations with the Agent
* **Interface Language**: Sets the display language of the Agent panel itself. Supports English, 简体中文, 日本語, 한국어, and Español

<img alt="SoloEnt Agent settings panel showing Preferred Language and Interface Language options" />

## Next steps

<CardGroup>
  <Card title="Theme settings" icon="palette" href="./theme-settings">
    Customize visual appearance
  </Card>

  <Card title="API Configuration" icon="code" href="./api-configuration">
    Configure API settings
  </Card>
</CardGroup>


# Login
Source: https://docs.soloent.ai/en/quickstart/login

Sign in with Google or Microsoft account

SoloEnt currently supports Google and Microsoft authentication for secure access.

<Info>
  More authentication methods coming soon.
</Info>

## Login steps

<Steps>
  <Step title="Launch SoloEnt">
    The login page opens automatically in your browser.
  </Step>

  <Step title="Choose login method">
    Select Google or Microsoft account.
  </Step>

  <Step title="Complete authentication">
    Sign in with your credentials. You may need to complete two-factor authentication.
  </Step>

  <Step title="Return to app">
    Click **"Open SoloEnt.app"** on the success page.
  </Step>

  <Step title="Authorize">
    Click **Open** in the SoloEnt authorization prompt.
  </Step>
</Steps>

## Troubleshooting

<AccordionGroup>
  <Accordion title="Browser doesn't open">
    Check if a browser window opened in the background, or manually copy the login URL from the SoloEnt window.

    * For more troubleshooting, see the FAQ: [The login redirect prompt does not appear](/en/faq#the-login-redirect-prompt-does-not-appear)
  </Accordion>

  <Accordion title="Authentication fails">
    * Verify your credentials
    * Try a different login method (Google vs Microsoft)
    * Check your internet connection
    * Disable VPN temporarily
  </Accordion>

  <Accordion title="Can't return to app">
    * Ensure SoloEnt is still running
    * Allow browser to open external applications
    * Restart SoloEnt and try again
  </Accordion>
</AccordionGroup>

<Tip>
  **For users in China:** Microsoft Account typically has better connectivity.
</Tip>

## Next steps

<CardGroup>
  <Card title="API configuration" icon="key" href="./api-configuration">
    Set up your AI model keys
  </Card>

  <Card title="Language settings" icon="language" href="./language-settings">
    Choose your preferred language
  </Card>

  <Card title="Theme settings" icon="palette" href="./theme-settings">
    Customize visual appearance
  </Card>

  <Card title="Start writing" icon="pen-to-square" href="../tips/create-project">
    Create your first project
  </Card>
</CardGroup>


# Quick Start
Source: https://docs.soloent.ai/en/quickstart/quickstart

Get up and running with SoloEnt in 5 minutes

Get started with SoloEnt by following these steps. You'll be writing with AI assistance in minutes.

## Setup steps

<Steps>
  <Step title="Launch and login">
    Open SoloEnt and sign in with your Google or Microsoft account. The login page opens automatically in your browser.

    <Card title="Login guide" icon="right-to-bracket" href="./login">
      Detailed login instructions
    </Card>
  </Step>

  <Step title="Choose preferences">
    Set your interface language and color theme.

    <CardGroup>
      <Card title="Language" icon="language" href="./language-settings">
        Interface and Agent language
      </Card>

      <Card title="Theme" icon="palette" href="./theme-settings">
        Color theme for comfort
      </Card>
    </CardGroup>
  </Step>

  <Step title="Choose your model">
    After logging in, you can start writing immediately using SoloEnt's official free models — no API key required.

    To access more models, configure a third-party API key in settings or upgrade your subscription to unlock VIP models.

    <Card title="Models & API configuration" icon="key" href="./api-configuration">
      Official free models, VIP subscription models, and third-party API key setup
    </Card>
  </Step>

  <Step title="Create your first project">
    Start a new writing project and begin creating content.

    <Card title="Create project" icon="folder-plus" href="../tips/create-project">
      Project creation guide
    </Card>
  </Step>
</Steps>

## Understanding the interface

SoloEnt uses a three-panel layout optimized for AI-assisted writing. Each panel serves a specific purpose in your workflow.

<Card title="Interface Overview" icon="grid-2" href="./interface">
  Explore the three-panel workspace in detail
</Card>

## Next steps

Choose your learning path:

<CardGroup>
  <Card title="Interface overview" icon="grid-2" href="./interface">
    Learn the workspace layout
  </Card>

  <Card title="Log in" icon="right-to-bracket" href="./login">
    Sign in to get started
  </Card>

  <Card title="API configuration" icon="key" href="./api-configuration">
    Set up your AI model keys
  </Card>

  <Card title="Start your first book" icon="pen-to-square" href="../tips/create-project">
    Create a new writing project
  </Card>
</CardGroup>


# Theme Settings
Source: https://docs.soloent.ai/en/quickstart/theme-settings

Choose eye-friendly color themes for comfortable writing

Select a color theme that's comfortable for your eyes during long writing sessions.

<Tip>
  Choose themes based on your environment and time of day to reduce eye strain.
</Tip>

## Choosing a theme

### On first launch

Select your initial theme from the welcome screen.

<img alt="SoloEnt theme settings menu" />

### Changing theme anytime

<Steps>
  <Step title="Open menu">
    Click **SoloEnt** (top-left) → **Settings**.
  </Step>

  <Step title="Go to Themes">
    Select **Themes** → **Color Theme**.
  </Step>

  <Step title="Preview and select">
    Use arrow keys (↑/↓) to preview themes. Press **Enter** to apply.
  </Step>
</Steps>

<img alt="SoloEnt theme settings menu" />

## Quick selection guide

| Environment             | Recommended Theme        |
| ----------------------- | ------------------------ |
| Bright room, daytime    | Light Default            |
| Evening, moderate light | Warm Light               |
| Night, low light        | Dark Blue or Dark Warm   |
| Late night              | Dark Default             |
| Need maximum contrast   | High Contrast Light/Dark |

## Eye comfort tips

<AccordionGroup>
  <Accordion title="Match your environment">
    Use light themes in bright rooms, dark themes in dim lighting.
  </Accordion>

  <Accordion title="Reduce blue light">
    Choose Warm Light or Dark Warm themes for evening writing.
  </Accordion>

  <Accordion title="Take breaks">
    Follow the 20-20-20 rule: every 20 minutes, look at something 20 feet away for 20 seconds.
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup>
  <Card title="API Configuration" icon="code" href="./api-configuration">
    Configure API settings
  </Card>

  <Card title="Start writing" icon="pen-to-square" href="../tips/create-project">
    Create your first project
  </Card>
</CardGroup>


# AI API Directory
Source: https://docs.soloent.ai/en/resources/api-directory

Complete guide to AI model providers and API registration

## Choosing a provider

Consider these factors:

**Budget:**

* **Tight budget** → DeepSeek, Google Gemini Flash, or [free options](./free-api-keys)
* **Moderate budget** → GPT-4, Claude Haiku
* **Premium quality** → GPT-5, Claude Sonnet, Claude Opus

**Use case:**

* **General writing** → OpenAI GPT-5.2
* **Strategic planning** → Claude Sonnet-4.5
* **Chinese content** → Zhipu GLM-5, DeepSeek-V3.2
* **Speed priority** → Google Gemini 3 Flash

## Popular AI providers

<Note>
  Prices listed here are subject to change. Please refer to the official provider for the most accurate pricing.
</Note>

### OpenAI

<Card title="OpenAI Platform" icon="openai" href="https://platform.openai.com">
  GPT-5, GPT-4, and more GPT models
</Card>

**Pricing:** Pay-as-you-go, starts at \$0.03/1K tokens for GPT-3.5

**How to register:**

<Steps>
  <Step title="Create account">
    Visit [platform.openai.com](https://platform.openai.com) and sign up
  </Step>

  <Step title="Add payment method">
    Navigate to Billing and add a payment method
  </Step>

  <Step title="Generate API key">
    Go to API keys section and create a new key
  </Step>
</Steps>

### Anthropic

<Card title="Anthropic Console" icon="microchip" href="https://console.anthropic.com">
  Claude Sonnet, Opus, and Haiku models
</Card>

**Pricing:** Pay-as-you-go, starts at \$3/million tokens for Haiku

### Google AI

<Card title="Google AI Studio" icon="google" href="https://aistudio.google.com">
  Gemini Pro and Flash models
</Card>

**Pricing:** Free tier available, Pro starts at \$0.35/million tokens

### DeepSeek

<Card title="DeepSeek Platform" icon="brain" href="https://platform.deepseek.com">
  DeepSeek V3 and R1 models
</Card>

**Pricing:** Very affordable, starts at ¥0.1/million tokens

### Zhipu AI

<Card title="Zhipu AI Open Platform" icon="language" href="https://open.bigmodel.cn">
  GLM-4, GLM-5 models
</Card>

**Pricing:** Competitive rates for Chinese market

## API aggregators

Aggregators provide access to multiple models through a single API key:

| Aggregator      | Models Available   | Benefits               |
| --------------- | ------------------ | ---------------------- |
| **OpenRouter**  | 100+ models        | One key for all models |
| **Together AI** | Open-source models | Self-hosted options    |
| **Replicate**   | Community models   | Wide variety           |

<Tip>
  Aggregators simplify managing multiple providers but may add a small markup.
</Tip>

## Free resources

Get started without spending money:

<Card title="Free API keys" icon="gift" href="./free-api-keys">
  Community resources, free tiers, and trial credits
</Card>

## Configuration help

After getting your API key, configure it in SoloEnt:

<Card title="API configuration guide" icon="key" href="../quickstart/api-configuration">
  Step-by-step setup instructions
</Card>

## Next steps

<CardGroup>
  <Card title="Configure API keys" icon="key" href="../quickstart/api-configuration">
    Set up your chosen provider
  </Card>

  <Card title="Free API keys" icon="gift" href="./free-api-keys">
    Start with free options
  </Card>

  <Card title="Start writing" icon="pen-to-square" href="../tips/create-project">
    Create your first project
  </Card>
</CardGroup>


# Free API Keys
Source: https://docs.soloent.ai/en/resources/free-api-keys

Get free AI model API keys for SoloEnt

Access AI models for free through community resources and provider free tiers.

## Community resources

<Card title="Join Discord Community" icon="users" href="https://discord.com/invite/NUh3skEFuY">
  Get updates on free API key giveaways and community shared keys
</Card>

Join the SoloEnt Discord community to receive:

* Free API key updates
* Community-shared trial keys
* Tips for maximizing free tier usage
* Help from other users

## Provider free tiers

### NVIDIA free tier

NVIDIA offers free access to select models through their developer platform:

<Card title="NVIDIA Build" icon="gift" href="https://build.nvidia.com">
  Free access to GLM-5 and Kimi-2.5 models
</Card>

**Available models:**

* GLM-5 (Zhipu AI)
* Kimi-2.5 (Moonshot)

<Note>
  NVIDIA free tier includes usage limits. Check their documentation for current quotas.
</Note>

**How to get started:**

<Steps>
  <Step title="Register account">
    Sign up at [build.nvidia.com](https://build.nvidia.com)
  </Step>

  <Step title="Get API key">
    Navigate to your dashboard and generate a free API key
  </Step>

  <Step title="Configure in SoloEnt">
    Add the key in SoloEnt's [API configuration](../quickstart/api-configuration)
  </Step>
</Steps>

<Warning>
  Trial credit policies change frequently. Always check the provider's official website for the most current offers and eligibility requirements.
</Warning>

<Card title="Complete API Directory" icon="book-open" href="./api-directory">
  View all providers with detailed registration guides and pricing
</Card>

## Limitations

Free API keys typically have:

* **Rate limits** - Fewer requests per minute
* **Usage quotas** - Monthly token caps
* **Feature restrictions** - May lack advanced features

<Info>
  For production use, consider upgrading to a paid plan for higher limits and better reliability.
</Info>

## Next steps

<CardGroup>
  <Card title="API directory" icon="book-open" href="./api-directory">
    Browse all AI providers
  </Card>

  <Card title="API configuration" icon="key" href="../quickstart/api-configuration">
    Configure your API keys
  </Card>

  <Card title="Start writing" icon="pen-to-square" href="../tips/create-project">
    Create your first project
  </Card>
</CardGroup>


# Local LLM Setup
Source: https://docs.soloent.ai/en/resources/local-llms

Run AI models on your own computer for privacy and offline access

<Note>
  This guide about Local LLMs was created 100% by AI. Please adjust based on your actual usage and requirements.
</Note>

## Why use local models?

<CardGroup>
  <Card title="Privacy" icon="shield">
    Your data never leaves your computer
  </Card>

  <Card title="No API costs" icon="dollar-sign">
    One-time setup, unlimited usage
  </Card>

  <Card title="Offline access" icon="wifi-slash">
    Work without internet connection
  </Card>

  <Card title="Full control" icon="sliders">
    Customize model parameters
  </Card>
</CardGroup>

## Requirements

<Warning>
  Local models require significant hardware resources. Recommended minimum:

  * **RAM**: 16GB (32GB+ for larger models)
  * **Storage**: 10-50GB per model
  * **GPU**: Optional but highly recommended (NVIDIA with 8GB+ VRAM)
</Warning>

## Popular solutions

### Ollama (Recommended for beginners)

Easy-to-use local model runner with simple CLI.

**Pros:**

* Simple installation
* Automatic model management
* Active community
* Optimized for Apple Silicon

**Best models:**

* Llama 3.1 (8B, 70B)
* Qwen 2.5
* DeepSeek Coder

<Card title="Ollama setup guide" icon="rocket" href="https://ollama.ai">
  Download and installation instructions
</Card>

### LM Studio

Desktop application with GUI for running local models.

**Pros:**

* User-friendly interface
* Model discovery and download
* Cross-platform (Mac, Windows, Linux)
* Built-in chat interface

**Best for:**

* Users who prefer GUI over CLI
* Testing multiple models easily
* Quick model comparison

<Card title="LM Studio" icon="desktop" href="https://lmstudio.ai">
  Download LM Studio
</Card>

### vLLM (Advanced)

High-performance inference engine for production deployments.

**Pros:**

* Fastest inference speed
* GPU optimization
* Production-ready
* API server included

**Best for:**

* Technical users
* High-throughput needs
* Custom deployments

## Connecting to SoloEnt

All local solutions expose an OpenAI-compatible API:

<Steps>
  <Step title="Start local server">
    Launch your chosen solution (Ollama, LM Studio, etc.)
  </Step>

  <Step title="Configure in SoloEnt">
    Use OpenAI-compatible configuration:

    ```plaintext theme={null}
    Base URL: http://localhost:11434/v1  (Ollama default)
    API Key: ollama  (or leave blank)
    Model ID: llama3.1  (your model name)
    ```
  </Step>

  <Step title="Test connection">
    Send a test message to verify the setup.
  </Step>
</Steps>

## Recommended models by use case

### Writing & storytelling

| Model             | Size | RAM Required | Quality   |
| ----------------- | ---- | ------------ | --------- |
| **Llama 3.1 70B** | 40GB | 64GB+        | Excellent |
| **Qwen 2.5 32B**  | 20GB | 32GB+        | Very good |
| **Llama 3.1 8B**  | 5GB  | 16GB+        | Good      |

### Chinese content

| Model            | Size | RAM Required | Quality   |
| ---------------- | ---- | ------------ | --------- |
| **Qwen 2.5 72B** | 42GB | 64GB+        | Excellent |
| **GLM-4 9B**     | 6GB  | 16GB+        | Very good |
| **DeepSeek 67B** | 38GB | 64GB+        | Excellent |

### Code & technical

| Model                 | Size | RAM Required | Quality   |
| --------------------- | ---- | ------------ | --------- |
| **DeepSeek Coder V2** | 16GB | 32GB+        | Excellent |
| **CodeLlama 34B**     | 20GB | 32GB+        | Very good |
| **Qwen 2.5 Coder 7B** | 4GB  | 8GB+         | Good      |

## Performance optimization

<AccordionGroup>
  <Accordion title="Use GPU acceleration">
    NVIDIA GPUs dramatically improve inference speed. Ensure CUDA is properly installed.
  </Accordion>

  <Accordion title="Choose appropriate model size">
    Larger models ≠ always better. 7B-13B models often provide the best speed/quality balance.
  </Accordion>

  <Accordion title="Quantization">
    Use Q4 or Q5 quantized models to reduce memory usage with minimal quality loss.
  </Accordion>

  <Accordion title="Adjust context length">
    Shorter context windows (4K-8K) run faster than long context (32K+).
  </Accordion>
</AccordionGroup>

## Common issues

<AccordionGroup>
  <Accordion title="Model runs too slowly">
    * Switch to smaller model (8B instead of 70B)
    * Use quantized version (Q4\_K\_M)
    * Enable GPU acceleration
    * Reduce context window size
  </Accordion>

  <Accordion title="Out of memory">
    * Choose smaller model
    * Close other applications
    * Upgrade RAM
    * Use higher quantization (Q3, Q4)
  </Accordion>

  <Accordion title="Connection refused">
    * Verify local server is running
    * Check Base URL and port number
    * Ensure no firewall blocking
    * Try [http://127.0.0.1](http://127.0.0.1) instead of localhost
  </Accordion>

  <Accordion title="Poor output quality">
    * Try different prompt format
    * Adjust temperature/top\_p settings
    * Switch to larger or different model
    * Check if model is appropriate for your language
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup>
  <Card title="API directory" icon="book-open" href="./api-directory">
    Browse cloud API providers
  </Card>

  <Card title="Free API keys" icon="gift" href="./free-api-keys">
    Get free cloud API access
  </Card>
</CardGroup>


# Markdown syntax
Source: https://docs.soloent.ai/en/resources/markdown

Text, title, and styling in standard markdown

## Titles

Best used for section headers.

```md theme={null}
## Titles
```

### Subtitles

Best used for subsection headers.

```md theme={null}
### Subtitles
```

<Tip>
  Each **title** and **subtitle** creates an anchor and also shows up on the table of contents on the right.
</Tip>

## Text formatting

We support most markdown formatting. Simply add `**`, `_`, or `~` around text to format it.

| Style         | How to write it   | Result            |
| ------------- | ----------------- | ----------------- |
| Bold          | `**bold**`        | **bold**          |
| Italic        | `_italic_`        | *italic*          |
| Strikethrough | `~strikethrough~` | ~~strikethrough~~ |

You can combine these. For example, write `**_bold and italic_**` to get ***bold and italic*** text.

You need to use HTML to write superscript and subscript text. That is, add `<sup>` or `<sub>` around your text.

| Text Size   | How to write it          | Result                 |
| ----------- | ------------------------ | ---------------------- |
| Superscript | `<sup>superscript</sup>` | <sup>superscript</sup> |
| Subscript   | `<sub>subscript</sub>`   | <sub>subscript</sub>   |

## Linking to pages

You can add a link by wrapping text in `[]()`. You would write `[link to google](https://google.com)` to [link to google](https://google.com).

Links to pages in your docs need to be root-relative. Basically, you should include the entire folder path. For example, `[link to text](/writing-content/text)` links to the page "Text" in our components section.

Relative links like `[link to text](../text)` will open slower because we cannot optimize them as easily.

## Blockquotes

### Singleline

To create a blockquote, add a `>` in front of a paragraph.

> Dorothy followed her through many of the beautiful rooms in her castle.

```md theme={null}
> Dorothy followed her through many of the beautiful rooms in her castle.
```

### Multiline

> Dorothy followed her through many of the beautiful rooms in her castle.
>
> The Witch bade her clean the pots and kettles and sweep the floor and keep the fire fed with wood.

```md theme={null}
> Dorothy followed her through many of the beautiful rooms in her castle.
>
> The Witch bade her clean the pots and kettles and sweep the floor and keep the fire fed with wood.
```

### LaTeX

Mintlify supports [LaTeX](https://www.latex-project.org) through the Latex component.

<Latex>8 x (vk x H1 - H2) = (0,1)</Latex>

```md theme={null}
<Latex>8 x (vk x H1 - H2) = (0,1)</Latex>
```


# Choose your plan
Source: https://docs.soloent.ai/en/subscription/choose-your-plan

Compare SoloEnt plans and select the one that fits your creative workflow

SoloEnt offers simple, transparent pricing. All plans include BYOK (Bring Your Own Key) support so you can use your existing API keys alongside SoloEnt AI.

## View pricing

<Card title="SoloEnt Pricing" icon="external-link" href="https://soloent.ai/en/pricing">
  Compare plans, features, and pricing on our pricing page
</Card>

## Plan overview

| Plan     | Price    | Models unlocked                                                                    |
| -------- | -------- | ---------------------------------------------------------------------------------- |
| **Free** | \$0      | Free-group LLMs; free BYOK                                                         |
| **Lite** | \$9.9/mo | VIP-group LLMs; free BYOK; tier-limited skill packs; member-only tools             |
| **Pro**  | \$20/mo  | VIP-group LLMs; free BYOK; tier-limited skill packs; member-only tools             |
| **Max**  | \$60/mo  | VIP-group LLMs; free BYOK; tier-limited skill packs; member-only tools; beta perks |

<Note>
  Pay for 10 months, get 12 months of features and AI budget (save 2 months) with yearly billing.
</Note>

Paid plans include a monthly SoloEnt AI usage allowance matching the subscription price. Subscriptions renew automatically at the chosen interval (monthly or annual) unless canceled. Payment is processed by our designated payment provider.

## What's included

* **SoloEnt AI** — Built-in AI, ready to use. Sign in and go—no API key setup required.
* **BYOK** — Use your existing API keys from OpenAI, Anthropic, Google, DeepSeek, Doubao, and more.
* **All plans** — Access to the listed VIP-group LLMs and product features.
* **Tier-limited** — Official skills and member-only tools (such as data, image, audio, and video services).

<Tip>
  Model availability may change during upgrades or promotional periods—see [https://soloent.ai/en/models](https://soloent.ai/en/models) or join [Discord](https://discord.com/invite/NUh3skEFuY) to stay updated on the latest offers.
</Tip>

## Unlimited add-on top-ups

On top of a Lite, Pro, or Max subscription, you can **top up extra credits based on your own usage — with no cap**.

<Note>
  Top-up credits are valid **only within the current billing cycle**. When your subscription renews next month or you make a subscription change (upgrade, downgrade, or switch billing period), your quota resets to the plan's base amount, and **top-up credits do not carry over**.
</Note>

For how to top up, see [Make the payment](./make-the-payment).

## Next steps

<Card title="Make the payment" icon="wallet" href="./make-the-payment">
  Complete your subscription with Stripe
</Card>


# Make the payment
Source: https://docs.soloent.ai/en/subscription/make-the-payment

Complete your SoloEnt subscription securely with Stripe

SoloEnt uses [Stripe](https://stripe.com) as its payment processor. Stripe acts as the Merchant of Record (MoR), handling checkout, payment processing, and compliance.

## Payment process

<Steps>
  <Step title="Go to your dashboard">
    There are two ways to access the subscription dashboard:

    * **Option 1**: Click your account avatar in the top-right corner of the Agent to open the menu, then select **Dashboard**
    * **Option 2**: Click **Upgrade subscription** at the bottom of the free model list

    Both options take you to [account.soloent.ai/en/dashboard](https://account.soloent.ai/en/dashboard/).

    <img alt="SoloEnt dashboard entry" />
  </Step>

  <Step title="Open the subscription plan tab">
    In the dashboard, navigate to **Subscription Plan** in the left sidebar to view available plans. Choose your plan (Lite, Pro, or Max) and billing cycle (monthly or yearly).

    <img alt="Subscription plan tab" />
  </Step>

  <Step title="Proceed to checkout">
    Click **Change plan** to start the checkout process. You will be redirected to Stripe's secure checkout page.
  </Step>

  <Step title="Enter payment details">
    Enter your payment information on the checkout page. Major credit cards and other regional payment methods are supported; Alipay subscriptions are supported via Stripe.

    <img alt="Stripe checkout page supporting Alipay and bank cards" />
  </Step>

  <Step title="Complete payment">
    Review your order and confirm. You will receive a confirmation email and your subscription will be activated.
  </Step>

  <Step title="Return to SoloEnt">
    After payment, you will be redirected back to SoloEnt. Your new plan features are available immediately.
  </Step>
</Steps>

## Security tips

<Warning>
  Never share your payment details, API keys, or account credentials with anyone. SoloEnt and Stripe will never ask for your full card number or password via email or chat.
</Warning>

**When making payments:**

* **Verify the URL** — Ensure you are on a legitimate domain (soloent.ai or stripe.com) before entering payment details.
* **Use secure connections** — Check for the padlock icon (HTTPS) in your browser's address bar.
* **Avoid public Wi‑Fi** — Prefer a private or trusted network when completing checkout.
* **Monitor your statements** — Review your bank or card statements to confirm charges match your subscription.
* **Keep credentials private** — Do not share your SoloEnt account password or session tokens.

<Info>
  Stripe uses industry-standard encryption and acts as the Merchant of Record, handling fraud prevention and payment compliance. API keys and webhooks are secured with authentication and HMAC-SHA256 signing.
</Info>

## Unlimited top-up credits

Subscribers on Lite, Pro, or Max plans can purchase additional credits at any time once their monthly quota is used up — there is no cap on top-ups.

<img alt="Unlimited top-up credits" />

<Warning>
  Top-up credits apply to the current billing cycle only. They do not carry over to the next cycle, and are reset when your subscription renews or if you change plans (upgrade, downgrade, or switch billing period).
</Warning>

## Troubleshooting

**Payment declined**
Check that your card details are correct and that your card supports international transactions. Some banks block overseas payments by default — contact your bank to authorize the charge, then retry.

**Subscription not activated after payment**
Wait a few minutes and refresh your dashboard. If your plan has not updated, contact support with your payment confirmation email.

**Duplicate charge**
If you see more than one charge for the same period, contact support immediately with your transaction IDs.

## Customer service

For billing issues, payment questions, or account problems, reach the SoloEnt team at:

<Card title="Contact support" icon="envelope" href="mailto:team@soloent.ai">
  [team@soloent.ai](mailto:team@soloent.ai)
</Card>

When contacting support, include:

* Your account email
* A description of the issue
* Your payment confirmation number or transaction ID (if applicable)

<Tip>
  For fastest response, join the [Discord community](https://discord.com/invite/NUh3skEFuY) where the team is active.
</Tip>

## Next steps

<CardGroup>
  <Card title="Choose your plan" icon="credit-card" href="./choose-your-plan">
    Compare plans and pricing
  </Card>

  <Card title="Manage subscription" icon="gear" href="./manage-subscription">
    Balance, invoices, and cancellation
  </Card>
</CardGroup>


# Manage subscription
Source: https://docs.soloent.ai/en/subscription/manage-subscription

Check your balance and invoices, and cancel your SoloEnt subscription

## Subscription rules

<AccordionGroup>
  <Accordion title="Cancellation">
    You may cancel a paid subscription at any time from your account or the subscription management page. After cancellation, you retain access until the end of the current billing period. No further charges will be made, and the subscription ends—access and AI allowance cease at that time.
  </Accordion>

  <Accordion title="Resuming">
    If you have chosen to cancel at period end, you may resume the subscription from your account before the period ends. After resuming, the subscription will continue to renew automatically.
  </Accordion>

  <Accordion title="Changing plans">
    After completing payment for a new plan, the new subscription takes effect and the previous one ends. AI budget and usage spend are reset according to the new plan, and billing continues at the chosen interval.
  </Accordion>

  <Accordion title="Refunds">
    If a refund is granted, that subscription will end and access and AI allowance will cease. Refund amount and timing depend on the payment provider's rules.
  </Accordion>
</AccordionGroup>

<Note>
  We may change plan features or pricing with reasonable advance notice. Such changes will not reduce your existing benefits during the current period unless otherwise stated. Specific plan features, pricing, and billing terms are described on our [pricing page](./choose-your-plan) and in your account dashboard.
</Note>

## Check your balance and invoice

Once payment is complete, your dashboard updates immediately to show your new balance and renewal date.

<img alt="Balance and renewal date in dashboard" />

You can also check your balance at any time in the client by opening your account info and refreshing.

<img alt="Balance in the client" />

In the dashboard, you can view your billing history.

<img alt="Billing history (invoice) 1 in the dashboard" />

<img alt="Billing history (invoice) 2 in the dashboard" />

## Cancel your subscription

In the dashboard, the button for your active plan changes to **Cancel subscription**. Click it and confirm when prompted—your current plan will be cancelled.

<img alt="Cancel subscription in dashboard" />

<Warning>
  After cancellation, you retain access until the end of the current billing period. When the period ends, access and AI allowance cease and the plan reverts to Free. No further charges will be made.
</Warning>

## Refund policy

Paid subscriptions are generally non-refundable; no refunds are given for billing periods already used.

If you have a reasonable refund request due to service failure, duplicate charges, or our error, please contact us at [team@soloent.ai](mailto:team@soloent.ai). We will evaluate your case and work with the payment provider where appropriate.

<Warning>
  If a refund is granted, that subscription will end immediately—access and AI allowance will cease. Refund amount and timing depend on the payment provider's rules.
</Warning>

## Customer service

For billing issues or account questions, reach the SoloEnt team at:

<Card title="Contact support" icon="envelope" href="mailto:team@soloent.ai">
  [team@soloent.ai](mailto:team@soloent.ai)
</Card>

When contacting support, include:

* Your account email
* A description of the issue
* Your payment confirmation number or transaction ID (if applicable)

<Tip>
  For fastest response, join the [Discord community](https://discord.com/invite/NUh3skEFuY) where the team is active.
</Tip>

## Next steps

<CardGroup>
  <Card title="Choose your plan" icon="credit-card" href="./choose-your-plan">
    Compare plans and pricing
  </Card>

  <Card title="Make the payment" icon="wallet" href="./make-the-payment">
    Payment process and security tips
  </Card>
</CardGroup>


# Save on Tokens
Source: https://docs.soloent.ai/en/subscription/save-tokens

How to use tokens efficiently and lower your monthly SoloEnt bill

## Prerequisite: Keep your client up to date

We continuously improve context management and API caching in every release, so that as models evolve we can keep delivering the best cost-efficiency for you.

<Card title="Download the latest version" icon="download" href="https://soloent.ai">
  Get the latest SoloEnt client from our website
</Card>

## The core equation

<Note>
  **Token usage = input size × number of calls**
</Note>

Once you internalize this, the playbook becomes simple: shrink each input and cut wasted calls.

## High impact — apply every session

### 1. Tighten the context window

Only show the AI what it actually needs. When you're writing chapter 47, it doesn't need chapter 1. When you're polishing one line of dialogue, it doesn't need the whole chapter.

**What to do**:

* Activate only the documents relevant to the current scene. When drafting a chapter, load only the directly relevant settings, the chapter outline, and limited context
* Maintain a [`SoloEnt.md`](../tips/SoloEnt) so the AI can absorb context from a single file instead of pulling in many
* Use `@` for precise references, or hold `shift` and drag specific files into the chat — don't open or read everything by default
* When editing dialogue, select only the target paragraph, not the entire chapter
* Close unused document references after each scene

<Tip>
  Estimated savings: **40–60%**
</Tip>

### 2. Replace long prose with short directives

The AI doesn't need your background framing — only what to do and how to do it. SoloEnt already provides the system prompt; you don't need to repeat the setup in chat.

**Token-heavy**:

```text theme={null}
You are a professional novel-writing assistant. Please rewrite this dialogue
to feel more tense, so the reader senses the strain between the two characters,
while keeping each character's voice consistent…
```

**Token-light**:

```text theme={null}
Rewrite dialogue: increase tension, preserve voice
```

Save your recurring directives as a [Skill](../tips/skills) — one click, zero descriptive cost.

<Tip>
  Estimated savings: **20–35%**
</Tip>

### 3. Audit the Rules you're loading

[Rules](../tips/rules) are the most overlooked silent token sink — they're force-loaded on every request.

**Trim them**:

* Load chapter-writing Rules only when actually writing chapters
* Delete "You are…" role-play preambles (the AI already knows what it is)
* Use lists instead of paragraphs — same information, half the tokens
* Audit Rules quarterly and remove anything the AI has already internalized

<Tip>
  Estimated savings: **15–30%**
</Tip>

## Medium impact — build good daily habits

### 4. Light tasks deserve light models

Not every task needs the strongest model.

| Task type                                                            | Best model (when quality matters) | Light model (when you can trade quality) |
| -------------------------------------------------------------------- | --------------------------------- | ---------------------------------------- |
| Brainstorming, outline generation, consistency checks                | Sonnet                            | Haiku, GLM                               |
| Prose writing, dialogue polish, scene expansion                      | Gemini                            | Doubao, DeepSeek                         |
| Complex plot design, deep style imitation, long-form logic threading | Opus                              | Sonnet, GLM                              |
| First-draft generation, outline drafting                             | GLM, DeepSeek                     | Open-source models                       |

<Tip>
  Estimated savings: **50–70%** on light-task workloads
</Tip>

### 5. Work in steps; don't ask for the full output in one shot

Don't probe by regenerating: asking for a 2,000-word chapter and restarting whenever you don't like it is **the most wasteful pattern there is**.

**Recommended flow** (chapter writing example):

<Steps>
  <Step title="Outline first">
    Have the AI produce the chapter structure and beats
  </Step>

  <Step title="Then expand">
    Once the outline is right, draft the prose
  </Step>

  <Step title="Tone and style pass">
    Polish locally at the end
  </Step>
</Steps>

Each step costs few tokens, and each one only continues after you've confirmed direction — total spend is far below repeated full regenerations.

**Use [Plan mode](../tips/plan-mode)**: before executing, switch to Plan mode and align on direction, structure, and key details over a few lightweight turns. Then switch back to execute. Plan mode burns very few tokens, and one round of alignment saves enormous spend on repeated regeneration.

```text theme={null}
[Plan mode]
This chapter has A and B reconciling, but I want to plant a seed for C.
What structures could work?
→ Align on direction and beats

[Execute mode]
Write the prose using structure 2
```

<Tip>
  Estimated savings: **30–50%** on iterative work
</Tip>

### 6. Open new windows often; don't keep extending old chats

Every chat window carries its history — the longer you talk, the bigger every subsequent input becomes, because the full history is replayed. A window that's run for dozens of turns can spend most of its budget on "historical baggage" alone.

**Suggestions**:

* After finishing one self-contained task, open a new window for the next
* Don't polish dialogue, discuss outlines, and edit settings in the same window
* If a window has grown long and you need to regenerate, prefer a fresh window with only the necessary context
* Re-activate the right context by referencing [`SoloEnt.md`](../tips/SoloEnt) or `@` specific files

<Note>
  Good habit: **one window, one job**
</Note>

<Tip>
  Estimated savings: **10–30%** over time
</Tip>

### 7. Tell the AI to edit, not rewrite

Without constraints, the AI tends to re-emit the whole passage. So **explicitly tell it what to change**.

**Triggers a full rewrite**:

```text theme={null}
Improve this passage
```

**Edits only**:

```text theme={null}
Only change paragraph 3 — slow the pacing of the sentences. Output only the
revised paragraph; nothing else.
```

Add "no explanation" / "no summary" — preambles and post-ambles cost tokens too.

<Tip>
  Estimated savings: **20–40%** on polish work
</Tip>

## Advanced — deeper optimization

### 8. Codify high-frequency flows as Workflows

If every chapter you write begins with the same ritual — review the previous summary, confirm character emotions, read the chapter outline — turn it into a [Workflow](../tips/workflows). The only parameter is the chapter number; everything else is assembled automatically.

The prompt tokens per call become a fixed minimum instead of a randomly inflated value, and consistency improves at the same time.

<Note>
  Outcome: **consistency + token savings**
</Note>

### 9. Use a local model as the "draft layer"

Run an open-source model locally with [LM Studio](../resources/local-llms) to produce the first draft (marginal cost: zero). Then use the cloud model for one final polish pass — small token spend, large quality lift.

**Hardware reference**:

| RAM   | Model size     | Suitable for        |
| ----- | -------------- | ------------------- |
| 16 GB | 7B parameters  | Drafting            |
| 32 GB | 13B parameters | More stable quality |

For prolific writers this can cut cloud spend by **60% or more**.

## In one sentence

<Note>
  Control the context and state your need precisely — don't over-engineer the prompt. That's the core of saving tokens.
</Note>

Short Rules, precise references, the right model for the task — do all three and your monthly token bill can drop by more than half, with no loss in writing quality.

## Next steps

<CardGroup>
  <Card title="Choose your plan" icon="credit-card" href="./choose-your-plan">
    Compare plans and pricing
  </Card>

  <Card title="Manage subscription" icon="gear" href="./manage-subscription">
    Balance, invoices, and cancellation
  </Card>
</CardGroup>


# Editor
Source: https://docs.soloent.ai/en/tips/Editor

Edit documents with SoloEnt Markdown editor

SoloEnt editor is a standard Markdown editor that supports all editing features for Markdown files.

<Info>
  Other file formats may need to be converted to Markdown first. See [file format conversion](/essentials/file-format-conversion).
</Info>

## Markdown syntax

The Markdown editor supports standard syntax for text styling, which you can preview in preview mode.

For detailed syntax reference, see Markdown common syntax.

### Common formatting

| Style         | Syntax            | Result            |
| ------------- | ----------------- | ----------------- |
| Bold          | `**bold**`        | **bold**          |
| Italic        | `_italic_`        | *italic*          |
| Strikethrough | `~strikethrough~` | ~~strikethrough~~ |
| Code          | `` `code` ``      | `code`            |

## Preview mode

Click the preview button in the top-right corner of the editor to view your document.

<img alt="SoloEnt editor with preview mode showing side-by-side view" />

<Tip>
  Preview mode helps you see the final rendering of your document in real-time.
</Tip>

## Split view

Click the split view button in the top-right corner of the editor to enable multi-pane display.

<img alt="Split view editor showing multiple panes" />

<Tip>
  Each click adds a new editor pane. Close all documents in a pane to remove it.
</Tip>

Split view allows you to view and edit multiple documents simultaneously, improving productivity.

## Office plugin

Users accustomed to Microsoft Word can install the "**Office Viewer**" plugin in SoloEnt for an Office-like editing experience.

<img alt="Office Viewer plugin showing Word-like editing interface" />

## Next steps

<CardGroup>
  <Card title="Create project" icon="folder-plus" href="./create-project">
    Start creating your first project
  </Card>

  <Card title="Plan mode" icon="lightbulb" href="./plan-mode">
    Organize your thoughts with plan mode
  </Card>

  <Card title="File format conversion" icon="rotate" href="./file-format-conversion">
    Convert different file formats
  </Card>

  <Card title="Markdown syntax" icon="text-size" href="../resources/markdown">
    Learn complete Markdown syntax
  </Card>
</CardGroup>


# SoloEnt.md
Source: https://docs.soloent.ai/en/tips/SoloEnt

The living memory system that keeps your writing project on track

`SOLOENT.md` is the **Central Control Panel** for writing projects in SoloEnt — the AI's long-term memory that records your story's DNA, tracks progress, and maintains consistency across tens of thousands of words.

It's **optional**, designed for medium to long-form writing. Shorter tasks may not need it — enable it when your project calls for it.

<Note>
  Like `CLAUDE.md` in coding projects, `SOLOENT.md` is the instruction manual for AI. It tells the AI what this project is about, how to write it, and where you are in the journey.
</Note>

## When to use SOLOENT.md

**Good fit:**

* Medium to long-form writing (novels, series, multi-chapter stories)
* Projects spanning multiple writing sessions that need consistency
* Stories with complex world-building, multiple characters, or careful foreshadowing

**Not necessary for:**

* Short stories or tasks completed in a single session
* Lightweight writing that doesn't require cross-session context tracking

***

Long-form writing faces a unique challenge: **context drift**. The AI might forget:

* The magic system rules you established in Chapter 1
* A character's motivation from three chapters ago
* The foreshadowing you carefully planted
* The tone and style you've been building

`SOLOENT.md` solves this by serving as the **Single Source of Truth** for your entire project.

## The 8 Core Sections

Every `SOLOENT.md` is structured into 8 sections, each with a specific purpose and lifecycle.

### 1. Project DNA (Foundation)

<Accordion title="Lifecycle: Static (Initialized once)">
  This section rarely changes after initialization. It defines the soul of your project.
</Accordion>

Defines the core identity of your project:

* **Genre**: Sci-fi, Romance, Mystery, etc.
* **Core Themes**: What the story is really about (e.g., "redemption," "power corrupts")
* **Target Audience**: Who will read this?
* **Emotional Promise**: What feeling should readers take away?
* **Word Count / Milestone Plan**: Scope and structure

**Example:**

```markdown theme={null}
## 1. Project DNA (Foundation)
- **Genre**: Psychological Horror / Supernatural
- **Core themes**: Paranoia, gaslighting, isolation, unreliable memory
- **Target audience**: Adult horror readers (18+)
- **Emotional promise**: A suffocating descent into madness
- **Word count / milestone plan**: ~20,000 words, 8-10 chapters
```

### 2. World System (Rules & Index)

<Accordion title="Lifecycle: Semi-static (Updated as world expands)">
  Add new locations, rules, or systems as your world grows. Core rules should remain stable.
</Accordion>

Records the rules of your fictional world to prevent logical conflicts:

* **Core World Rules**: Magic systems, technology levels, social structures
* **World Index**: Links to detailed world-building documents

**Example:**

```markdown theme={null}
## 2. World System (Rules & Index)

### 2.1 Core World Rules (Summary)
> Reality is unstable within the Building. Small changes accumulate. 
> Memory cannot be trusted. Physical evidence disappears or transforms.

### 2.2 World Index (Pointers)
- [The Building] -> `world/the-building.md`
- [Magic System] -> `world/magic-system.md`
```

<Tip>
  Use index pointers for complex settings. Keep `SOLOENT.md` concise by storing detailed world-building in separate files.
</Tip>

### 3. Characters (Index & Relations)

<Accordion title="Lifecycle: Semi-static (Updated as characters are added)">
  Add new characters as they appear. Update relationships when dynamics shift significantly.
</Accordion>

Tracks not just names, but the dynamic state and relationships of characters:

* **Character Index**: Main and supporting cast with links to detailed profiles
* **Relationship Network**: Visual or text-based relationship map

**Example:**

```markdown theme={null}
## 3. Characters (Index & Relations)

### 3.1 Character Index (Concise)
- **Mei Chen** (32, Translator, Protagonist) -> `characters/mei-chen.md`
- **The Landlord** (Age unknown, Building superintendent)

### 3.2 Relationship Network (Summary)
- Mei Chen → The Landlord: Polite distance, growing unease
- Mei Chen → Previous Tenant: Parasitic connection, identity bleed
```

### 4. Plot & Structure (Index)

<Accordion title="Lifecycle: Semi-static (Updated if outline changes)">
  The overall structure is stable, but you may adjust chapter breakdowns as you write.
</Accordion>

Provides macro-level control of the story flow:

* **Story Structure**: Three-Act, Hero's Journey, or custom structure
* **Outline Index**: Links to detailed chapter outlines

**Example:**

```markdown theme={null}
## 4. Plot & Structure (Index)

### 4.1 Story Structure
- **Act I (Ch. 1-3)**: Establishment - Mei moves in, small wrongness begins
- **Act II (Ch. 4-6)**: Escalation - Reality fractures
- **Act III (Ch. 7-8)**: Climax - The horrifying truth emerges

### 4.2 Outline Index (Pointers)
- [Master Outline] -> `outline/master-outline.md`
- [Chapter Breakdown 01] -> `outline/chapter-breakdown-01.md`
```

### 5. Stylistic Guidelines (Deep Deconstruction)

<Accordion title="Lifecycle: Static (Unless style shifts)">
  Set this once and adjust only if you need to course-correct the AI's writing style.
</Accordion>

The **most powerful section** for tuning AI writing style:

* **Base Tone**: Overall atmosphere (e.g., "Claustrophobic, oppressive")
* **Language & Rhetoric**: Sentence structure, word choice, literary devices
* **Dialogue Style**: How characters speak
* **Pacing Rules**: Chapter rhythm and tension management

**Example:**

```markdown theme={null}
## 5. Stylistic Guidelines (Deep Deconstruction)

- **Base Tone**: Claustrophobic, oppressive, quietly unhinged
- **Language & Rhetoric**: Clean, restrained prose that fractures alongside 
  the protagonist's mind. Short sentences during panic.
- **Dialogue Style**: Sparse. Conversations feel slightly off—people almost 
  say what they mean.
- **Pacing Rules**: Slow burn with sudden stabs of terror. Each chapter 
  ends on a hook.
```

<Warning>
  Be specific in this section. "Write beautifully" is vague. "Use short, punchy sentences under 15 words" is actionable.
</Warning>

### 6. Key Notes & Soft Constraints

<Accordion title="Lifecycle: Auto-Updated + User Updated (Hybrid)">
  **Auto-Updated**: The Agent learns from your feedback and revision requests, automatically extracting your preferences and constraints.

  **User Updated**: You can also manually add constraints anytime to steer the AI.
</Accordion>

Your direct instructions to the AI:

* **Narrative Taboos**: What NOT to write (e.g., "No romance subplot")
* **User Constraints**: Specific requirements you care about

<Tip>
  Use this section to fix recurring issues. If the AI keeps adding romance when you don't want it, add "No romantic subplots" to Narrative Taboos.
</Tip>

**Example:**

```markdown theme={null}
## 6. Key Notes & Soft Constraints

### Narrative Taboos
- No jump scares as primary horror device
- No excessive gore (psychological horror is the focus)
- Avoid explaining too much—ambiguity is the weapon

### User Constraints
- Psychological horror focus: paranoia, gaslighting, mind games
- Reader should question what's real alongside protagonist
```

<Warning>
  **🧠 Intelligent Learning System**

  This section has a special power: **it learns from your feedback automatically**.

  When you ask the Agent to revise something:

  * "Make the dialogue less formal" → Adds constraint: "Keep dialogue casual and natural"
  * "Remove excessive descriptions" → Adds taboo: "No purple prose or over-description"
  * "The pacing is too slow" → Adds guideline: "Tighten pacing, cut unnecessary scenes"

  Over time, `SOLOENT.md` becomes a personalized style guide that captures YOUR unique writing preferences.
</Warning>

### 7. Active Writing State (Dynamic Context)

<Accordion title="Lifecycle: Auto-Updated (After every writing session)">
  The Agent automatically updates this section as you write. You rarely need to touch it manually.
</Accordion>

The **most dynamic section**, updated by the Agent after every writing session:

* **Current Status**: Where are we in the story?
* **Character Snapshot**: Current physical/mental state of characters
* **Recent Plot Flow**: Summary of what just happened
* **Pending Threads**: Open plot hooks and foreshadowing to remember
* **Execution Review**: Quality check for the latest chapter

**Example:**

```markdown theme={null}
## 7. Active Writing State (Dynamic Context)

### 7.1 Current Status
- **Current Chapter**: Chapter 2 - "Moving In" (Completed)
- **Story Time**: November 1st, 3:17 AM (first night in apartment)
- **Current Location**: Apartment 4C, bedroom
- **Immediate Atmosphere**: Terror, sleeplessness, questioning sanity

### 7.2 Character Snapshot
- **Mei Chen**: Terrified, sleep-deprived, finding notes in her own 
  handwriting, fearing she's inherited her grandmother's mental illness

### 7.3 Recent Plot Flow
- **Ch.1**: Mei viewed 4C, met Mr. Kadir, signed lease
- **Ch.2**: Moving day—elevator broken, lost time, duplicate cup appeared

### 7.4 Pending Threads
- **Urgent Hooks**: The note from her dream, the duplicate cups
- **Active Foreshadowing**: The closet door that won't latch
- **Next Chapter Goal**: The notes continue and escalate
```

<Check>
  This section is why `SOLOENT.md` works. It gives the AI perfect memory of where you left off.
</Check>

### 8. Project Roadmap & Milestones

<Accordion title="Lifecycle: Auto-Updated (As tasks complete)">
  The Agent checks off tasks and logs milestones as you progress. You can manually add new tasks anytime.
</Accordion>

Tracks the project's progress and next steps:

* **Current Focus**: What is the immediate task?
* **Todo List**: Checklist of upcoming chapters or tasks
* **Milestone Log**: History of completed major goals

**Example:**

```markdown theme={null}
## 8. Project Roadmap & Milestones

### 8.1 Current Focus & Next Steps
- **Phase**: First Draft Sprint
- **Core Task**: Writing Chapter 3 - "The Notes"
- **Todo List**:
  - [x] Initialize SOLOENT.md
  - [x] Write Chapter 1
  - [x] Write Chapter 2
  - [ ] Write Chapter 3
  - [ ] Write Chapters 4-9

### 8.2 Milestone Log
- [Completed] Project initialized with psychological horror focus
- [Completed] Chapter 1 completed - Established unsettling atmosphere
- [Completed] Chapter 2 completed - First major horror beats delivered
```

## How SOLOENT.md is Generated

### Manual Initialization (/init)

`SOLOENT.md` **is not generated automatically**. You can create it at any stage of your project using the `/init` command:

<Steps>
  <Step title="Type /init in the chat">
    Send the `/init` command at any point in a writing session, regardless of the project stage.
  </Step>

  <Step title="Agent gathers project info">
    The Agent organizes information about genre, themes, characters, and plot — based on existing content or by asking questions.
  </Step>

  <Step title="SOLOENT.md is generated">
    Based on the current project state, `SOLOENT.md` is created with all 8 sections and saved in the project root.
  </Step>
</Steps>

<Tip>
  You don't have to create it at the start. Running `/init` later — once you feel the project needs context tracking — works just as well.
</Tip>

### Automatic Updates

Once `SOLOENT.md` exists, the Agent automatically maintains it as you write:

| Trigger                | What Gets Updated                                                                  |
| ---------------------- | ---------------------------------------------------------------------------------- |
| **Finish a chapter**   | Section 7 (Active Writing State) - Recent Plot Flow, Character Snapshot            |
| **Complete a task**    | Section 8 (Project Roadmap) - Todo List, Milestone Log                             |
| **Add new character**  | Section 3 (Characters) - Character Index                                           |
| **Establish new rule** | Section 2 (World System) - Core World Rules                                        |
| **Request revisions**  | Section 6 (Key Notes & Constraints) - Learns your preferences and adds constraints |

<Info>
  The Agent updates `SOLOENT.md` silently in the background. Once the file exists, you don't need to ask for updates — they happen automatically.
</Info>

## How to Maintain SOLOENT.md

### Agent Maintenance (Automatic)

The Agent handles routine updates:

* ✅ Summarizing completed chapters
* ✅ Tracking pending plot threads
* ✅ Updating character states
* ✅ Checking off completed tasks

### User Maintenance (Manual)

You control the creative direction:

<CardGroup>
  <Card title="Refine Style" icon="palette">
    Edit **Section 5** to adjust tone, pacing, or dialogue style if the AI's writing feels off.
  </Card>

  <Card title="Set Boundaries" icon="shield">
    Update **Section 6** to add narrative taboos or constraints when the story drifts.
  </Card>

  <Card title="Steer Direction" icon="compass">
    Modify **Section 7.4 (Pending Threads)** to prioritize certain plot lines.
  </Card>

  <Card title="Add New Ideas" icon="lightbulb">
    Insert new characters, locations, or rules into **Sections 2-3** as inspiration strikes.
  </Card>
</CardGroup>

## Best Practices

### 1. Keep it Fresh

<Warning>
  A stale `SOLOENT.md` leads to poor AI performance. If Section 7 (Active Writing State) is outdated, the AI will write as if you're still in Chapter 2 when you're actually on Chapter 8.
</Warning>

**How to keep it fresh:**

* Let the Agent update it after each chapter
* Review Section 7 at the start of each writing session
* Manually sync if you skip chapters or rewrite sections

### 2. Be Explicit About What You Don't Want

The AI is eager to please, which sometimes means adding things you didn't ask for.

**Use Section 6 (Narrative Taboos) to set hard boundaries:**

```markdown theme={null}
### Narrative Taboos
- No sudden superpowers for the protagonist
- No deus ex machina rescues
- No romantic subplot
- No excessive exposition dumps
```

### 3. Use Indexes for Complex Settings

Don't bloat `SOLOENT.md` with 10 pages of magic system details.

**Instead, use pointers:**

```markdown theme={null}
### 2.2 World Index (Pointers)
- [Magic System] -> `world/magic-system.md`
- [Political Structure] -> `world/politics.md`
- [Geography] -> `world/map.md`
```

The Agent can read these files when needed, keeping the main control panel clean.

### 4. Load Context at Session Start

<Tip>
  At the beginning of each writing session, mention `SOLOENT.md` in your prompt (e.g., "Let's continue writing. Check SOLOENT.md for context."). This ensures the AI loads the latest state.
</Tip>

### 5. Review Before Major Decisions

Before writing a climactic chapter or making a major plot decision:

1. Read through `SOLOENT.md` yourself
2. Check if pending threads from Section 7.4 need resolution
3. Verify character states in Section 7.2 match your intent

## Maintenance Workflow

<Steps>
  <Step title="Write a chapter">
    Work with the Agent to draft your chapter.
  </Step>

  <Step title="Agent auto-updates">
    After completion, the Agent updates Section 7 (Active Writing State) and Section 8 (Roadmap).
  </Step>

  <Step title="Review updates">
    Quickly scan the updates to ensure accuracy.
  </Step>

  <Step title="Manual adjustments (optional)">
    If you want to steer the next chapter differently, edit Section 7.4 (Pending Threads) or Section 6 (Constraints).
  </Step>

  <Step title="Start next session">
    Begin your next writing session by referencing `SOLOENT.md` for context.
  </Step>
</Steps>

## Update Frequency by Section

| Section                        | Update Frequency          | Who Updates                             |
| ------------------------------ | ------------------------- | --------------------------------------- |
| **1. Project DNA**             | Once at initialization    | User (rarely changes)                   |
| **2. World System**            | When world expands        | Agent + User                            |
| **3. Characters**              | When characters are added | Agent + User                            |
| **4. Plot & Structure**        | If outline changes        | User                                    |
| **5. Stylistic Guidelines**    | If style needs correction | User                                    |
| **6. Key Notes & Constraints** | After revisions + Anytime | **Agent (learns from feedback) + User** |
| **7. Active Writing State**    | After every chapter       | Agent (auto)                            |
| **8. Project Roadmap**         | As tasks complete         | Agent (auto)                            |

## Common Pitfalls

<AccordionGroup>
  <Accordion title="Letting Section 7 get outdated">
    **Problem**: You write 3 chapters without updating Active Writing State.

    **Result**: The AI forgets recent plot developments and writes inconsistently.

    **Solution**: Let the Agent update after each chapter, or manually sync before starting a new session.
  </Accordion>

  <Accordion title="Vague style guidelines">
    **Problem**: Section 5 says "Write beautifully" or "Make it exciting."

    **Result**: The AI doesn't know what you mean and defaults to generic prose.

    **Solution**: Be specific. "Use sentences under 15 words. Focus on tactile sensory details. Avoid adverbs."
  </Accordion>

  <Accordion title="Forgetting to use indexes">
    **Problem**: You paste 20 pages of world-building directly into `SOLOENT.md`.

    **Result**: The file becomes bloated and hard to navigate.

    **Solution**: Keep summaries in `SOLOENT.md`, store details in linked files.
  </Accordion>

  <Accordion title="Not setting narrative taboos">
    **Problem**: You don't explicitly state what you don't want.

    **Result**: The AI adds a romance subplot you hate, or resolves tension too quickly.

    **Solution**: Use Section 6 to clearly state "No romance," "No easy solutions," etc.
  </Accordion>
</AccordionGroup>

## Advanced Tips

### For Multi-POV Stories

Create character snapshots for each POV in Section 7.2:

```markdown theme={null}
### 7.2 Character Snapshot
- **Protagonist A** (POV 1): Currently in the castle, unaware of the betrayal
- **Protagonist B** (POV 2): Just discovered the truth, planning revenge
```

### For Non-Linear Narratives

Track story time vs. narrative time in Section 7.1:

```markdown theme={null}
### 7.1 Current Status
- **Current Chapter**: Chapter 5 (Flashback)
- **Story Time**: 1995 (20 years before main timeline)
- **Narrative Time**: Revealing why the protagonist fears water
```

### For Series or Sequels

Link to previous books' `SOLOENT.md` files:

```markdown theme={null}
## 1. Project DNA (Foundation)
- **Series**: Book 2 of "The Dark Trilogy"
- **Previous Book**: [Book 1 SOLOENT.md] -> `../book-1/SOLOENT.md`
```

## More Advanced Tips

<CardGroup>
  <Card title="Plan Mode" icon="map" href="./plan-mode">
    Use Plan Mode to design your story structure before writing
  </Card>

  <Card title="Commands" icon="terminal" href="./commands">
    Powerful workflows at your fingertips with slash commands
  </Card>
</CardGroup>


# Agent Image Generation
Source: https://docs.soloent.ai/en/tips/agent-image

Have the Agent generate covers, character cards, and illustrations directly from natural language — no model switching required

<Note>
  Image tools are available from **v0.11.0** and are **members-only**.
</Note>

From v0.11.0, the Agent supports an image-generation tool. Once enabled, you can chat with the Agent as usual and complete image tasks with natural language.

## Where to enable

In **Settings**, you can turn **SoloEnt Image Tools** on or off.

<img alt="The SoloEnt Image Tools toggle in the TOOLS section of Settings, described as generating covers and illustrations with AI" />

Once enabled, note that:

* You complete image tasks by **talking to the Agent in natural language**
* **No model switching is required** — run the skill with the model you normally use, and image generation calls the service automatically
* The image service is **provided officially**, automatically consumes your account's AI credits, and is **members-only**

## Make covers or character cards with a Skill

With a writing Skill, you can have the Agent generate covers, character cards, or illustrations directly from your novel's worldbuilding and character settings. The example below uses a "character card" Skill.

<Steps>
  <Step title="Install the character-card Skill">
    Install the "character card" Skill in your project (available in one click from the [Marketplace](/en/tips/marketplace)).

    <img alt="Installing the character-card Skill in the project" />
  </Step>

  <Step title="Make your request in chat">
    Following the Skill's usage, make your request in the chat window, for example "Generate a character card for the female lead."

    * The Skill automatically locates and reads the files related to that character's settings and confirms the details with you
    * Pick a model you normally use with **strong execution ability**; this model does not affect the image generation

    <img alt="Asking the Agent in the chat window to generate a character card for the female lead" />

    After you answer all of the Skill's guided questions, it produces an image **generation prompt** (an example is in the collapsed block below).
  </Step>

  <Step title="Confirm and generate">
    Once the prompt matches the character card you have in mind, confirm generation and the Agent will automatically call the image tool.

    <img alt="After confirming the generation prompt, the Agent automatically calls the image tool" />
  </Step>

  <Step title="Collect the result">
    The generated image appears directly in your project folder.

    <img alt="The generated character-card image appears directly in the project folder" />
  </Step>
</Steps>

<Accordion title="Example: character-card generation prompt (Xiao Chen)">
  **Text character card**

  ```text theme={null}
  Name: Xiao Chen
  Age / Identity: 22 · unlicensed masseuse · Chen's Massage (Chinatown)
  Visual anchors: (1) white model label "SomaTouch Basic v2.3" on the back of the neck; (2) the edge of the grey concealing spray peels up, revealing real skin tone and faint freckles underneath; (3) slightly thin but strong hands with thin calluses on the palms (an android hand would have no calluses).
  Core desire: to be seen as a real person, rather than pretending to be a machine.
  Fatal wound: her true identity is Harrison Wells's illegitimate daughter — but she does not know it herself.
  Key prop: the neck model label (put on every day, peeled off every day — both disguise and shackle).
  Signature line: "Every day I pretend to be a machine. But I know I'm not."
  ```

  **Generation prompt** (submitted to the model)

  ```text theme={null}
  Character key visual illustration for a near-future dystopian sci-fi novel, restrained cold tone with hidden warmth, cinematic film still aesthetic, low-key lighting with film grain, like "Her" or "Arrival".
  Subject: 22-year-old female, unlicensed masseuse disguised as an android, quiet and restrained aura hiding stubborn defiance, black straight hair tucked under a cap showing only a standardized hairline, deep brown eyes (hollow and distant during work, warm when makeup removed), pale skin covered by Soma standard grey concealing spray.
  Signature traits: (1) white model label "SomaTouch Basic v2.3" on the back of her neck; (2) grey concealing spray on face and neck — one edge peeling up to reveal real pale skin with faint freckles underneath; (3) slender but powerful hands with thin calluses on palms (android hands don't have calluses).
  Costume: simple grey short-sleeve T-shirt, low neckline exposing the neck label, dark trousers, old canvas sneakers; cheap but clean, practical and unadorned.
  Pose and expression: full body standing in a dim massage parlor back room, body slightly turned away but face turned toward the viewer — like a candid moment caught on camera; one hand unconsciously touching the edge of the neck label, gaze restrained but pupils holding a trace of living warmth and unease; lips slightly pursed, not smiling, not angry, just quietly existing.
  Key prop: white model label "SomaTouch Basic v2.3" on the back of the neck, one corner peeling up to reveal real skin and freckles; an old massage table faintly visible behind her.
  Background: blurred dim underground massage parlor back room — warm yellow old bulb, peeling wallpaper, massage supplies stacked in corners; overall dark, a single side light from above illuminating the character's silhouette, creating a cinematic lonely atmosphere.
  Color: primary cold grey (Soma standard color / disguise) + secondary warm yellow (old lamp / true humanity) + accent deep brown (eyes), cold-warm contrast pointing to "humanity beneath a machine shell".
  Composition: full body standing, 9:16 vertical portrait, camera slightly below eye level (subtle low angle), main light from warm yellow lamp above and to the side, half-lit half-shadow, film grain texture.
  Style: cinematic film still aesthetic, low-key lighting, restrained mood like "Her" or "Arrival", film grain, naturalistic lighting, highly detailed.
  Aspect ratio: 9:16 portrait, 576x1024 pixels.
  On-card text: minimal text at bottom — small name "Xiao Chen" in thin white type on a translucent dark strip, no stat panel, no long quote, visual-first key visual treatment, subtle and integrated into the scene.
  Negative: no watermark, no garbled characters, no misspelled text, no unintended text, no extra limbs, no extra fingers, no over-designed text frames, no large text blocks.
  ```
</Accordion>

<img alt="An example character card generated by the Agent" />

<CardGroup>
  <Card title="Skills" icon="wand-magic-sparkles" href="/en/tips/skills">
    Learn how to create and use Skills, and pair them with image generation for covers and character cards
  </Card>

  <Card title="Marketplace" icon="store" href="/en/tips/marketplace">
    Install official writing skill packs in one click
  </Card>
</CardGroup>


# Commands
Source: https://docs.soloent.ai/en/tips/commands

Powerful workflows at your fingertips with slash commands

SoloEnt includes built-in **slash commands** that give you instant access to professional writing workflows. Just type `/` in the editor to invoke these specialized tools.

<Note>
  Think of commands like power tools in your workshop. Instead of manually checking every detail, you invoke a command and let the Agent run a comprehensive workflow for you.
</Note>

## How to Use Commands

<img alt="Commands interface" />

<Steps>
  <Step title="Type / in the editor">
    Press `/` (forward slash) to open the command palette.
  </Step>

  <Step title="Select a command">
    Choose from the list of available commands (e.g., `review`, `market`).
  </Step>

  <Step title="Agent runs the workflow">
    The Agent automatically executes the pre-configured workflow and generates a detailed report.
  </Step>

  <Step title="Review the results">
    The Agent saves the report to your `notes/` directory and provides actionable suggestions.
  </Step>
</Steps>

## Built-in Commands

SoloEnt currently provides four official commands. Here is what each does at a glance:

| Command        | What it does                                                                                                                   |
| -------------- | ------------------------------------------------------------------------------------------------------------------------------ |
| **`/init`**    | Manually initialize the project memory file `SOLOENT.md` — see the [SOLOENT.md guide](/en/tips/SoloEnt)                        |
| **`/compact`** | Manually compact the current session's context to save tokens — see [Using the Context Window Wisely](/en/tips/context-window) |
| **`/review`**  | Run a multi-layered quality review on a chapter (details below)                                                                |
| **`/market`**  | Analyze your work's publishing/marketing strategy (details below)                                                              |

### `/review` - Quality Assurance

The `/review` command runs a multi-layered quality assurance process on your chapter:

<AccordionGroup>
  <Accordion title="1. Basic Review">
    * **Word count**: Checks if chapter length matches your target
    * **Outline alignment**: Verifies the chapter follows your planned outline
  </Accordion>

  <Accordion title="2. Consistency Review">
    * **Story continuity**: Compares with story outline and act structure
    * **Chapter flow**: Checks consistency with previous chapter
    * **Character integrity**: Verifies characters act according to their profiles
    * **World rules**: Ensures no violations of established world-building
  </Accordion>

  <Accordion title="3. Writing Style Review">
    * **Style guide compliance**: Checks against your stylistic guidelines (Section 5 of SOLOENT.md)
    * **Character voice**: Ensures characters speak and act consistently
    * **Tone consistency**: Verifies the chapter maintains the intended atmosphere
  </Accordion>

  <Accordion title="4. Chapter Ending Review">
    * **Hook quality**: Ensures chapter ends with a compelling hook
    * **Ending pitfalls**: Flags repeated words, long inner monologues, or weak endings
    * **Next chapter connection**: Verifies smooth transition to next chapter
  </Accordion>

  <Accordion title="5. AI Writing Pattern Detection">
    * **Originality check**: Verifies text doesn't reproduce copyrighted material
    * **Repeated phrases**: Detects overused words or phrases
    * **Em dash overuse**: Flags unnecessary em dash usage
    * **List pattern abuse**: Catches "word, word, word" patterns common in AI writing
  </Accordion>
</AccordionGroup>

<Tip>
  We highly suggest you run `/review` after completing each chapter, before moving to the next. This catches issues early when they're easier to fix.
</Tip>

**Output:**

The Agent generates a detailed review report saved to:

```
notes/review-chapter-{number}-{date}.md
```

### `/market` - Publishing Strategy

The `/market` command analyzes your completed manuscript and generates a comprehensive publishing strategy:

<AccordionGroup>
  <Accordion title="1. Genre & Market Analysis">
    * **Primary genre**: Identifies your book's genre and subgenre
    * **Word count**: Calculates per-chapter and total word count
    * **Market category**: Determines MG/YA/Adult and Fiction/Non-fiction classification
  </Accordion>

  <Accordion title="2. Story Elements Extraction">
    * **Logline**: Crafts a one-sentence hook for your story
    * **Main selling points**: Identifies unique elements that make your book marketable
    * **Comparable titles**: Suggests "comp titles" for query letters
  </Accordion>

  <Accordion title="3. Publishing Route Recommendation">
    * **Best-fit route**: Traditional, Hybrid, Self-publishing, or Fandom platforms
    * **Rationale**: Explains why this route fits your manuscript
    * **Target platforms**: Specific publishers, imprints, or self-publishing platforms
    * **Next steps**: Query letter, editing needs, cover design, etc.
  </Accordion>

  <Accordion title="4. Success Benchmarks">
    * **Comparable titles**: Books similar to yours and their success metrics
    * **Timeline**: Expected timeline to publication
    * **Budget**: Estimated costs (for self-publishing or hybrid routes)
  </Accordion>
</AccordionGroup>

<Tip>
  Run `/market` when your manuscript is complete or near-complete. This helps you plan your publishing strategy before diving into the query process.
</Tip>

**Output:**

The Agent generates a marketing plan saved to:

```
notes/marketing-plan-{date}.md
```

## More Advanced Tips

<CardGroup>
  <Card title="SOLOENT.md Guide" icon="book-sparkles" href="./SoloEnt">
    Learn how commands interact with your project constitution
  </Card>

  <Card title="Plan Mode" icon="map" href="./plan-mode">
    Use Plan Mode to structure your story before writing
  </Card>
</CardGroup>


# Rules vs. Workflows vs. Skills
Source: https://docs.soloent.ai/en/tips/compare

Three ways to instruct your Agent—each does a different job. Learn the difference so your Agent actually works the way you do.

Rules, Workflows, and Skills are three different ways to give your Agent instructions, and they each handle a separate concern. Mix them up and you'll find your Agent forgetting your voice, skipping steps, or ignoring your world-building.

| Type         | What it does                         | When it applies                  | One-line analogy                  |
| ------------ | ------------------------------------ | -------------------------------- | --------------------------------- |
| **Rule**     | Sets hard limits and non-negotiables | Always, automatically            | Company policy handbook           |
| **Workflow** | Defines a step-by-step process       | When you trigger a specific task | A standard operating procedure    |
| **Skill**    | Packages a specialized technique     | When you call on it by name      | A trained specialist on your team |

## Rule

A Rule is a standing constraint. No matter which chapter you're writing or what you ask the Agent to do, it follows these rules automatically—you never have to remind it.

**What belongs here:**

* **Banned words and phrases**—things you never want to appear in your manuscript

  ```
  Never use: suddenly, very, needless to say, as if on cue
  No italics for internal monologue—write it as action or dialogue
  Don't end dialogue with "he/she said" as a beat tag
  ```

* **Character voice floors**—make sure your leads can't be written out of character

  ```
  Nathaniel: terse, never admits defeat in words—only through action.
              Won't say "I was wrong." Responds to pressure by going quiet.
  Vera: tough exterior, never shows vulnerability in front of strangers
  ```

* **World-building hard limits**—the rules of your fictional universe can't be broken

  ```
  Setting is 1890s London. No electricity, no motor vehicles.
  Magic system has five tiers—there is no sixth.
  The Order of Ashford never harms civilians. This is a core plot rule.
  ```

<Tip>
  If you have a clear set of lines you don't want the Agent to cross, start with Rules. They're the lowest-effort, highest-impact thing you can set up.
</Tip>

***

## Workflow

A Workflow is a procedure. You tell the Agent how to handle a type of task step by step—then whenever you need it, you call it and it runs the whole process without hand-holding.

**What belongs here:**

* **Chapter continuation**—a fixed sequence every time you continue a scene

  ```
  When continuing a chapter, follow these steps:
  1. Recap the emotional hook from the previous chapter's final beat
  2. Confirm the main conflict goal for this chapter
  3. Write the scene (target: 1,500–2,000 words)
  4. End on a cliffhanger or reversal
  ```

  How to use it: just say "continue chapter twelve" and the Agent handles all four steps—no need to remind it to end on a hook.

* **Manuscript review**—hand off a draft and get a structured edit report back

  ```
  On receiving a draft, check in order:
  1. Character name consistency throughout
  2. Timeline conflicts (e.g. daytime scene suddenly shifts to night)
  3. Dialogue that breaks character voice (cross-check against Rules)
  4. Output a report in this format: Issue / Paragraph / Suggested fix
  ```

* **Outline expansion**—take a one-line premise and build it into a full chapter structure

<Tip>
  If you keep explaining the same steps to the Agent over and over, turn it into a Workflow. This is where you'll feel the biggest time savings.
</Tip>

***

## Skill

A Skill is a reusable writing technique. You teach the Agent a specific craft move once, save it, and then call on it whenever you need it—without re-explaining how it works.

**What belongs here:**

* **Signature emotional beats**—lock in the pacing of a scene type you use often

  ```
  [Skill: slow-burn confession]
  - Open with the lead acting indifferent; the reader assumes he doesn't care
  - Plant one specific detail mid-scene that reveals he's been paying attention all along
  - Close with a single action or line of dialogue—no internal monologue—that makes his
    feelings undeniable
  - Pacing: long setup → one quiet detonation → white space ending
  ```

  How to use it: "Use the slow-burn confession skill for the scene where Caleb sees Mara at the train station."

* **Action scene style**—lock in how your fight scenes move and breathe

* **Voice matching**—save your own prose style so the Agent doesn't drift into generic AI phrasing

<Tip>
  If you have a craft technique that's hard to re-explain every time—a pacing move, a tonal register, a genre convention—put it in a Skill.
</Tip>

***

## Using all three together

Rules, Workflows, and Skills aren't in competition. They layer on top of each other.

**Example: writing a contemporary romance novel**

* **Rule**: Declan never apologizes first. Ban phrases like "her heart raced" and "butterflies in her stomach." No supernatural elements—this is strictly contemporary.
* **Workflow**: Before each continuation, confirm the previous chapter's emotional hook. After writing, check whether the main characters' dialogue stays in voice.
* **Skill**: Save a "Declan-style almost-confession" technique—he never says "I love you." He shows it through small, specific actions, and the reader has to put it together.

All three run at once. That's when the Agent starts writing something that actually sounds like your book.

***

## Where to start

<Steps>
  <Step title="Start with Rules">
    List the mistakes you most dread the Agent making—banned words, character voice floors, world-building limits. This is the easiest thing to set up and has an immediate effect.
  </Step>

  <Step title="Then build a Workflow">
    Pick your most repetitive task—chapter continuation or manuscript review—and make one Workflow for it. You'll feel the time savings right away.
  </Step>

  <Step title="Add Skills as you go">
    When you find yourself re-explaining the same craft technique, that's the signal to package it as a Skill. You don't need to build these all at once—add them as the need comes up.
  </Step>
</Steps>

***

## Further reading

<CardGroup>
  <Card title="Rules" icon="scroll" href="/en/tips/rules">
    Capture your personal writing preferences so the Agent always writes in your voice, not its default
  </Card>

  <Card title="Workflows" icon="diagram-project" href="/en/tips/workflows">
    Package repetitive writing tasks into reusable step files and run an entire process with a single command
  </Card>

  <Card title="Skills" icon="wand-magic-sparkles" href="/en/tips/skills">
    Package specialized Agent knowledge into reusable instruction sets that activate only when you need them
  </Card>
</CardGroup>


# Using the Context Window Wisely
Source: https://docs.soloent.ai/en/tips/context-window

Understand the context window's limits and costs, and use task isolation, session compaction, and doc offloading to keep long-form writing stable and token-efficient

## What is the context window?

Every session window has a context-length limit. Above the chat box you can see how much context the current window has used — for example the **30.2k / 300.0k** below: the left number is the used session size, and the right number is the current session's context ceiling.

<img alt="The context progress bar above the chat box, showing 30.2k used and a 300.0k ceiling" />

## Setting the context ceiling

A single session's context ceiling is determined by the model's capability; today's mainstream models already support up to 1M. SoloEnt lets you set the per-session context ceiling to a reasonable level as needed.

<img alt="The Context window dropdown in model settings, letting you choose between the 300K and 1M tiers" />

<Warning>
  **Bigger is not always better.** An oversized window scatters the model's attention — especially during repeated exploration and revision, performance can actually drop.
</Warning>

* **Keep it around 200K–300K.**
* Common practice: following mature Agent products like Codex Cloud, people usually use only the first tier (200K–300K) of context and rarely go higher.

## Why we don't recommend maxing out a single 1M window

<AccordionGroup>
  <Accordion title="Surging cost of one long-running session">
    Many users assume a single window brings caching advantages, but some models charge in **tiers based on session size**, so an overly long session risks a cost surge.

    <Note>
      Example: with some models, input/output below 256K is one price tier, and above 256K it rises noticeably — a clear tiered price jump.
    </Note>
  </Accordion>

  <Accordion title="Uncertainty of the cache">
    While most new models offer good caching (a cache hit may cost only 1/5 to 1/10 of the price), sessions **spanning days or long idle gaps** cause the cache to expire, so a single request consumes a lot of tokens.
  </Accordion>

  <Accordion title="Switching models = starting over with a huge context">
    If you switch models partway through a single window, the new model actually has to read all the context the old model accumulated, which can consume a large chunk of the session at once — very uneconomical. **Always summarize the prior session before switching models.**
  </Accordion>
</AccordionGroup>

## Best practices

Switching windows means giving up the current window's session history, which feels wasteful to many users; but over-relying on a single window is neither economical nor good for long-term context quality in practice.

* **Task isolation**: aim for "one session, one task," and check completion before moving on.
  * If the next task is related: summarize the current session, offload it to a project file (such as `SOLOENT.md`), then continue in a new session.
  * If the next task is unrelated: decisively open a new session to avoid piling up historical tokens.
* **Use session compaction**: when a session gets too full, use the `/compact` command to condense the current content into a short summary and use it as the starting input for a new session — preserving the core information while lowering token usage.

<img alt="The default-command hint shown when typing /compact in the chat box: compact the current session to free up context space" />

* Once context usage passes about **200K**, continuing may cause a cost surge; a good rule of thumb is to consider opening a new window when a single window reaches **100–150K**.

### Always summarize and offload before opening a new window

1. **Long-form fiction**: use `/init` to generate `SOLOENT.md` in your project and write key context — characters, outline, current progress — into it. Each new window lets the model quickly restore the full project picture by reading `SOLOENT.md`, rather than relying on context accumulated in a single session.
2. **After heavy discussion**: if you and the Agent have discussed a lot and context is running high, ask the Agent to summarize the current discussion before switching windows (or models) and land the conclusions in a readable document. After switching, read that document first to get back up to speed.

<CardGroup>
  <Card title="SOLOENT.md Guide" icon="file-lines" href="/en/tips/SoloEnt">
    Use a memory file to carry long-term context so every new session quickly restores the full picture
  </Card>

  <Card title="Commands" icon="terminal" href="/en/tips/commands">
    Learn how to use built-in commands like `/compact` and `/init`
  </Card>
</CardGroup>


# Create Your First Project
Source: https://docs.soloent.ai/en/tips/create-project

Set up a new writing project in SoloEnt

In SoloEnt, each work is an independent project stored as a folder on your local computer.

<Info>
  The Agent automatically creates and organizes files within your project folder as needed.
</Info>

## Creating a new project

<Steps>
  <Step title="Choose location">
    Select the parent folder where you want to create your project.
  </Step>

  <Step title="Create new folder">
    Click **New Folder** and name your project.
  </Step>

  <Step title="Open in SoloEnt">
    SoloEnt opens your new project folder automatically.
  </Step>
</Steps>

## Opening an existing project

<Steps>
  <Step title="Click Open">
    Select **Open Folder** in SoloEnt.
  </Step>

  <Step title="Navigate to project">
    Browse to your project folder location.
  </Step>

  <Step title="Select folder">
    Choose the project folder and click **Open**.
  </Step>
</Steps>

<Warning>
  Always open the **project folder**, not individual files. This allows the Agent to manage your entire project structure.
</Warning>

## Folder permissions

<Accordion title="Why does SoloEnt need folder access?" icon="lock">
  **First-time access:**

  * Your system may request your password for security
  * This grants SoloEnt permission to read and write files

  **Security:**

  * Permissions are limited to the specific folder you open
  * SoloEnt cannot access other parts of your system
  * You can revoke permissions at any time in system settings
</Accordion>

## Project organization tips

<CardGroup>
  <Card title="One project per folder" icon="folder">
    Keep each work in its own dedicated folder
  </Card>

  <Card title="Descriptive names" icon="tag">
    Use clear project names (e.g., "SciFi-Novel-2026")
  </Card>

  <Card title="Use version control" icon="code-branch">
    Track changes with timeline for version history
  </Card>

  <Card title="Avoid special characters" icon="circle-exclamation">
    Use letters, numbers, hyphens, and underscores only
  </Card>
</CardGroup>

## What happens after opening?

Once you open a project folder:

1. **File manager populates**: The left panel shows your project structure
2. **Agent activates**: Ready to help with your writing tasks
3. **Editor ready**: Start creating or editing files
4. **You're all set to get started**: Begin your creative journey

## Starting your first writing session

<Steps>
  <Step title="Tell the Agent your idea">
    Describe what you want to write in the Agent chat.

    Example: "I want to write a fantasy novel about a young wizard"
  </Step>

  <Step title="Let Agent set up structure">
    The Agent will create necessary folders and initial files.
  </Step>

  <Step title="Start writing">
    Begin working on your outline, characters, or first chapter.
  </Step>
</Steps>

## Troubleshooting

<AccordionGroup>
  <Accordion title="Can't open folder">
    * Check folder permissions
    * Ensure folder exists and path is correct
    * Try creating a new folder instead
    * Restart SoloEnt and try again
  </Accordion>

  <Accordion title="Permission denied">
    * Enter your system password when prompted
    * Check system security settings
    * Try a different folder location
    * Run SoloEnt with appropriate permissions
  </Accordion>

  <Accordion title="Files not showing">
    * Verify you opened a folder, not a file
    * Check if folder contains compatible files
    * Refresh the file manager
    * Restart SoloEnt
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup>
  <Card title="Plan Mode" icon="lightbulb" href="./plan-mode">
    Learn strategic planning with the Agent
  </Card>

  <Card title="Editor features" icon="pen-to-square" href="./Editor">
    Master the writing editor
  </Card>

  <Card title="Agent commands" icon="terminal" href="./commands">
    Use powerful writing commands
  </Card>

  <Card title="Methodology" icon="book-open" href="./methodology">
    Understand SoloEnt's writing approach
  </Card>
</CardGroup>


# File Format Conversion
Source: https://docs.soloent.ai/en/tips/file-format-conversion

Convert Word documents to Markdown format

Import your existing Word documents into SoloEnt by converting them to Markdown format.

## Single file conversion

Convert individual `.docx` files to Markdown:

<Steps>
  <Step title="Locate the file">
    Find your `.docx` file in the file manager.
  </Step>

  <Step title="Right-click">
    Right-click on the file to open the context menu.
  </Step>

  <Step title="Convert">
    Select **"Convert to Markdown"** from the menu.
  </Step>

  <Step title="Wait for completion">
    The file converts automatically. A new `.md` file appears in the same location.
  </Step>
</Steps>

<Info>
  The original `.docx` file remains unchanged. The converted Markdown file is created as a new file.
</Info>

## Batch conversion

Convert multiple Word documents at once using the “**Markdown Hub**” plugin:

<Steps>
  <Step title="Install Markdown Hub">
    1. Click the **Extensions** icon in the left sidebar (third icon)
    2. Search for **"Markdown Hub"**
    3. Click **Install** (it's free)
  </Step>

  <Step title="Organize files">
    Place all `.docx` files you want to convert in the same folder.
  </Step>

  <Step title="Right-click folder">
    Right-click on the folder containing your `.docx` files.
  </Step>

  <Step title="Convert all">
    Select **"Convert to Markdown"** to batch convert all documents in the folder.
  </Step>
</Steps>

<Tip>
  Use batch conversion to import your entire existing work library at once.
</Tip>

## Supported formats

| Format  | Import | Export | Notes          |
| ------- | ------ | ------ | -------------- |
| `.docx` | ✅      | ❌      | Word documents |
| `.md`   | ✅      | ✅      | Native format  |
| `.txt`  | ✅      | ✅      | Plain text     |
| `.pdf`  | ❌      | ✅      | Export only    |

<Note>
  Markdown (`.md`) is SoloEnt's native format. All editing features work best with Markdown files.
</Note>

## Conversion quality

### What converts well

* **Text content**: Paragraphs, headings, lists
* **Basic formatting**: Bold, italic, underline
* **Structure**: Chapters, sections, subsections

### What may need adjustment

* **Complex tables**: May require manual formatting
* **Images**: Need to be manually placed in the images folder
* **Custom styles**: Convert to standard Markdown formatting
* **Comments and tracked changes**: Not preserved

<Warning>
  Review converted files for formatting issues. Complex Word documents may need manual cleanup.
</Warning>

## After conversion

<Steps>
  <Step title="Review content">
    Open the converted Markdown file and check for formatting issues.
  </Step>

  <Step title="Fix formatting">
    Adjust any elements that didn't convert cleanly.
  </Step>

  <Step title="Organize files">
    Move files to appropriate folders (chapters, outline, etc.).
  </Step>

  <Step title="Start editing">
    Begin working with your imported content.
  </Step>
</Steps>

<Tip>
  **Let Agent do the work:** Ask Agent to handle these tasks automatically. For example: "Review the converted file and fix formatting issues" or "Organize imported files into chapters folder."
</Tip>

## Best practices

<AccordionGroup>
  <Accordion title="Prepare Word documents">
    * Use standard styles (Heading 1, 2, 3, etc.)
    * Avoid complex formatting
    * Remove unnecessary styling before conversion
  </Accordion>

  <Accordion title="Test with one file first">
    * Convert a single file to check quality
    * Adjust your Word document if needed
    * Then proceed with batch conversion
  </Accordion>

  <Accordion title="Backup originals">
    * Keep your original `.docx` files
    * Store backups before converting
    * You can always reconvert if needed
  </Accordion>
</AccordionGroup>

## Troubleshooting

<AccordionGroup>
  <Accordion title="Conversion fails">
    * Ensure file is a valid `.docx` format
    * Check file isn't corrupted or password-protected
    * Try opening and resaving in Word
    * Convert to `.txt` first as a workaround
  </Accordion>

  <Accordion title="Formatting looks wrong">
    * This is normal for complex documents
    * Manually adjust Markdown formatting
    * Simplify original Word document and reconvert
  </Accordion>

  <Accordion title="Can't find converted file">
    * Check the same folder as the original
    * Look for a file with the same name but `.md` extension
    * Refresh the file manager
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup>
  <Card title="Editor features" icon="pen-to-square" href="./Editor">
    Learn Markdown editing
  </Card>

  <Card title="Organize project" icon="folder-tree" href="./create-project">
    Structure your imported files
  </Card>
</CardGroup>


# Git Version Control
Source: https://docs.soloent.ai/en/tips/git

Initialize your project as a local Git repo, commit snapshots at key milestones, and roll back to a version you are happy with

SoloEnt lets you version-control your whole project (workspace) with Git. You can initialize the open project as a **local repository** (no internet or remote hosting required), commit staged progress as snapshots while writing, and later roll the entire project back to a satisfactory point if you are not happy with the AI's subsequent changes.

## Initialize the repository

In the left sidebar, select the third tab (**Source Control**) and click **Initialize Repository** to turn the current project into a local Git repo.

<img alt="The Source Control tab in the left sidebar, showing project files like chapter-01.md and the auto-commit entry available after initializing the repository" />

## Commit changes

After initializing, each time you finish a task you can commit the changes to the repository from the Source Control panel.

<Steps>
  <Step title="Review changes">
    The left side of the panel lists the files changed this time (new files are marked **U**).
  </Step>

  <Step title="Generate a commit message">
    Click the second icon in the top-right of the panel to auto-generate a commit message with SoloEnt's **commit summary** feature.
  </Step>

  <Step title="Commit">
    After confirming the message, click **Commit** to save this change as a version checkpoint.
  </Step>
</Steps>

<img alt="The Source Control commit view: the list of changed files on the left, and the auto icon next to the commit button in the top-right used to generate a commit message" />

## When to commit

If you are worried that a version's content might later be lost or messed up, commit after each **milestone task**, for example:

* Finishing the outline for the whole book
* Finishing the first 10 chapters
* Making a major worldbuilding or setting change

That way, if you later feel the AI's edits aren't good, you can roll the whole project back to a previously committed checkpoint and start over.

<Tip>
  The more frequently and cleanly you commit, the easier it is to find that clean "before it got messed up" version when rolling back.
</Tip>

<CardGroup>
  <Card title="Context & SOLOENT.md" icon="file-lines" href="/en/tips/SoloEnt">
    Use `SOLOENT.md` to record the full picture of your project so every new session can quickly restore context
  </Card>

  <Card title="FAQ" icon="circle-question" href="/en/faq">
    Check here for issues like the "install Git" prompt and checkpoints
  </Card>
</CardGroup>


# Marketplace
Source: https://docs.soloent.ai/en/tips/marketplace

Browse and install official writing plugins—Rules, Workflows, Skills, and combinations—with a single tap

The Marketplace is the official distribution channel for writing plugins. Each plugin may contain Rules, Workflows, Skills, or a combination of all three — designed for a specific writing scenario, such as a "romance continuation assistant" or a "chapter review workflow." No manual configuration needed; once installed, follow the plugin's documentation to start using it.

## Opening the Marketplace

Click the **Marketplace icon** at the top of the SoloEnt Agent panel to open the Plugin Marketplace. You can browse in two ways:

* Search by keyword using the search bar
* Use the language filter in the top-right corner to find plugins suited for a specific language

<img alt="Plugin Marketplace page showing All / Installed / Updates tabs, a search bar, and a language filter dropdown with English selected; options include 中文, 日本語, 한국어, and Español" />

Click a plugin to open its detail page. Each plugin detail shows:

* **Name and summary**: a sentence or two describing what problem this plugin solves
* **Installs**: how many users have installed this plugin
* **Usage**: how to invoke or trigger the plugin
* **Contents**: what combination of Rules, Workflows, and Skills it includes
* **Core rules**: a quick overview of the plugin's underlying logic
* **Version**: the currently available version number

<img alt="Plugin detail page showing the World Building plugin's description, Install button, author, version, install count, and README with usage instructions and included Workflows" />

<Card title="Browse all plugins" icon="store" href="https://soloent.ai/en/plugins">
  View community and official plugins on the SoloEnt website
</Card>

## Installing a plugin

<Warning>
  You must have a project open before installing. Plugins install into the currently active project by default, not your global directory.
</Warning>

<Steps>
  <Step title="Open your project">
    Before installing any plugin, open the writing project you want to use it in.
  </Step>

  <Step title="Choose a plugin">
    Browse or search the Marketplace, then click a plugin card to read its full documentation.
  </Step>

  <Step title="Click Install">
    Click the **Install** button on the plugin detail page. Once complete, the plugin appears in the **Installed** tab.
  </Step>
</Steps>

## Verifying the installation

After installation, check the following locations to confirm the files were written successfully:

| Content type | Where to look               |
| ------------ | --------------------------- |
| Rule         | Agent panel → Rules tab     |
| Workflow     | Agent panel → Workflows tab |
| Skill        | Agent panel → Skills tab    |

For combined plugins, files will appear in multiple locations.

<img alt="Workflows tab showing the three-layer-outline.md workflow file installed from the Marketplace, labeled as Plugin source" />

## Using a plugin

Each plugin's detail page includes usage instructions covering:

* **How to trigger it**: how to invoke the included Workflows or Skills
* **Prerequisites**: what you need to have ready before using (for example, an open chapter file)
* **Expected output**: what the plugin produces when it runs

Open the Agent panel and follow the trigger instructions in the documentation.

* **Rule** takes effect automatically after installation — no manual invocation needed
* **Workflow** is triggered using the corresponding command
* **Skill** activates automatically based on context

<Tip>
  After installing, read the plugin's full documentation before using it.
</Tip>

## Coming in future versions

A future release will allow users to upload their own plugins, making it possible to share the Rules, Workflows, and Skills you've refined with other writers.

<CardGroup>
  <Card title="Skills" icon="wand-magic-sparkles" href="/en/tips/skills">
    Learn how Skills work and create your own from scratch
  </Card>

  <Card title="Workflows" icon="diagram-project" href="/en/tips/workflows">
    Understand Workflow structure and customize your writing process
  </Card>
</CardGroup>


# Plan Mode
Source: https://docs.soloent.ai/en/tips/plan-mode

Understand when and how to use Plan Mode for strategic planning

Plan Mode helps you think through complex writing tasks before taking action. It's designed for situations where you need to explore ideas, clarify requirements, and create a strategy before making changes.

## Switching between modes

You can switch between Plan Mode and Act Mode using the mode selector in the bottom-left corner of the SoloEnt Agent chat window.

<img alt="Plan and Act mode selector in SoloEnt Agent" />

<CardGroup>
  <Card title="Plan Mode" icon="lightbulb">
    Think first, act later
  </Card>

  <Card title="Act Mode" icon="bolt">
    Execute tasks immediately
  </Card>
</CardGroup>

## Plan Mode: "Think before you act"

In Plan Mode, the AI's primary task is to understand your requirements and create a blueprint. It won't immediately modify files or execute commands. Instead, it will confirm with you: "Here's what I plan to do."

### When to use Plan Mode

<AccordionGroup>
  <Accordion title="Starting a new project with vague ideas">
    When you're beginning a new work with only a rough concept (for example: "I want to write a science fiction novel"), Plan Mode helps you explore and structure your ideas.
  </Accordion>

  <Accordion title="Facing complex writing challenges">
    When you encounter difficult writing situations that require external input and discussion (such as writer's block or uncertainty about plot progression).
  </Accordion>

  <Accordion title="Making significant structural changes">
    When you need to make large-scale changes that affect multiple files or require careful consideration of trade-offs.
  </Accordion>
</AccordionGroup>

### How Plan Mode works

<Steps>
  <Step title="Analyze the problem">
    The AI examines your request and identifies key challenges and requirements.
  </Step>

  <Step title="Break down the steps">
    It creates a detailed action plan with clear steps and rationale.
  </Step>

  <Step title="Request confirmation">
    Before taking any action, it presents the plan for your approval.
  </Step>

  <Step title="Refine as needed">
    You can discuss, modify, and improve the plan until you're satisfied.
  </Step>
</Steps>

### Key characteristics

<CardGroup>
  <Card title="Strategic thinking" icon="chess">
    Ideal for handling complex requirements that need multiple rounds of discussion and clarification.
  </Card>

  <Card title="High safety" icon="shield-check">
    Shows you the plan without modifying your documents.
  </Card>

  <Card title="Transparent reasoning" icon="brain">
    Explains why certain approaches are recommended, helping you identify gaps and improvements.
  </Card>

  <Card title="Collaborative" icon="users">
    Encourages dialogue and refinement before execution.
  </Card>
</CardGroup>

<Warning>
  In Plan Mode, SoloEnt Agent will not execute any tasks or generate any files. It only provides strategic guidance and planning.
</Warning>

## Act Mode: "Just do it"

Once you switch to Act Mode, the AI becomes an executor. It will directly perform operations based on established goals or identified steps.

### When to use Act Mode

Use Act Mode when your task is very clear and specific:

* "Generate an outline for the first 10 chapters"
* "Change the protagonist's name to Wang"
* "Create a character profile for the antagonist"
* During normal writing sessions when you know what you want

<Tip>
  For most regular writing tasks, Act Mode is the default choice. Only switch to Plan Mode when you need strategic thinking and planning.
</Tip>

### How Act Mode works

<Steps>
  <Step title="Identify the task">
    The AI recognizes what needs to be done based on your request.
  </Step>

  <Step title="Execute immediately">
    It performs the necessary actions without asking for confirmation.
  </Step>

  <Step title="Adjust and optimize">
    Based on your feedback, it refines or modifies the results.
  </Step>
</Steps>

### Key characteristics

<CardGroup>
  <Card title="High efficiency" icon="rocket">
    Skips repeated confirmations and delivers results directly.
  </Card>

  <Card title="Automated" icon="wand-magic-sparkles">
    Can automatically create files, modify content, and run commands.
  </Card>

  <Card title="Immediate feedback" icon="gauge-high">
    You see changes and results right away.
  </Card>

  <Card title="Action-oriented" icon="play">
    Focuses on execution rather than discussion.
  </Card>
</CardGroup>

<Info>
  If you notice your Agent is only providing suggestions without executing or generating files, check if you're in Plan Mode.
</Info>

## Choosing the right mode

Use this decision tree to select the appropriate mode:

```mermaid theme={null}
graph TD
    A[New task] --> B{Is the task clear and specific?}
    B -->|Yes| C[Use Act Mode]
    B -->|No| D{Do you need to explore options?}
    D -->|Yes| E[Use Plan Mode]
    D -->|No| F{Are there multiple approaches?}
    F -->|Yes| E
    F -->|No| C
    E --> G[Discuss and refine plan]
    G --> H{Plan approved?}
    H -->|Yes| I[Switch to Act Mode]
    H -->|No| G
    I --> C
    C --> J[Execute task]
```

## Best practices

<AccordionGroup>
  <Accordion title="Start with Plan Mode for new projects">
    When beginning a new writing project, use Plan Mode to establish structure, characters, and plot before diving into writing.
  </Accordion>

  <Accordion title="Switch to Act Mode for execution">
    Once you have a clear plan, switch to Act Mode to implement it efficiently.
  </Accordion>

  <Accordion title="Return to Plan Mode when stuck">
    If you encounter writer's block or need to make significant changes, switch back to Plan Mode to think through options.
  </Accordion>

  <Accordion title="Use Plan Mode for major revisions">
    When restructuring your work or making large-scale changes, Plan Mode helps you avoid mistakes and maintain consistency.
  </Accordion>
</AccordionGroup>

## Configure different models for different modes

Official models let you quickly configure different models for **Plan Mode** and **Act Mode**:

* Some models are more organized during planning and improve back-and-forth clarification, such as Sonnet 4.6 and GLM-5;
* Some models are closer to what execution needs — more concise, or lower cost — such as Gemini and DeepSeek;

Setting them separately isn't required; it suits power users who understand different models deeply.

<img alt="The API Configuration panel showing Plan Mode and Act Mode tabs, with the 'Use different models for Plan and Act modes' option at the bottom" />

## Common scenarios

### Scenario 1: Starting a new novel

<Steps>
  <Step title="Use Plan Mode">
    Discuss genre, themes, characters, and plot structure.
  </Step>

  <Step title="Review and refine">
    Iterate on the plan until you're satisfied with the direction.
  </Step>

  <Step title="Switch to Act Mode">
    Begin writing chapters based on the approved plan.
  </Step>
</Steps>

### Scenario 2: Fixing a plot hole

<Steps>
  <Step title="Switch to Plan Mode">
    Analyze the plot hole and explore potential solutions.
  </Step>

  <Step title="Evaluate options">
    Discuss pros and cons of different approaches.
  </Step>

  <Step title="Choose a solution">
    Select the best approach for your story.
  </Step>

  <Step title="Switch to Act Mode">
    Implement the chosen solution.
  </Step>
</Steps>

### Scenario 3: Daily writing session

<Steps>
  <Step title="Stay in Act Mode">
    For routine writing tasks, Act Mode is more efficient.
  </Step>

  <Step title="Write continuously">
    Focus on producing content without interruption.
  </Step>

  <Step title="Switch to Plan Mode if needed">
    Only switch if you encounter a major decision point.
  </Step>
</Steps>

## Troubleshooting

<AccordionGroup>
  <Accordion title="Agent not executing tasks">
    **Problem:** The Agent only provides suggestions without taking action.

    **Solution:** Check if you're in Plan Mode. Switch to Act Mode to enable execution.
  </Accordion>

  <Accordion title="Agent executing without confirmation">
    **Problem:** The Agent makes changes without asking for approval.

    **Solution:** You're in Act Mode. Switch to Plan Mode if you want to review plans before execution.
  </Accordion>

  <Accordion title="Planning takes too long">
    **Problem:** The Agent spends too much time planning.

    **Solution:** If your task is straightforward, switch to Act Mode for faster results.
  </Accordion>

  <Accordion title="Need to modify the plan">
    **Problem:** The proposed plan doesn't match your vision.

    **Solution:** Stay in Plan Mode and continue the discussion. Provide more context or constraints to refine the plan.
  </Accordion>
</AccordionGroup>

## Next steps

<CardGroup>
  <Card title="Create a new project" icon="file-plus" href="./create-project">
    Learn how to start a new writing project with SoloEnt
  </Card>

  <Card title="Editor features" icon="pen-to-square" href="./Editor">
    Explore the writing and editing capabilities
  </Card>

  <Card title="Methodology" icon="book-open" href="./methodology">
    Understand SoloEnt's writing methodology
  </Card>

  <Card title="Commands" icon="terminal" href="./commands">
    Master the command system for efficient writing
  </Card>
</CardGroup>


# Rules
Source: https://docs.soloent.ai/en/tips/rules

Capture your personal writing preferences so the Agent always writes in your voice, not its default

Rules are markdown files that give the Agent persistent writing instructions. Define them once, and the Agent follows them automatically — across every session, every chapter, every prompt.

## Two types of rules

### Global rules

Global rules apply to every writing project you open. Use them for personal habits that you carry from project to project:

* Your natural sentence rhythm preferences
* Punctuation style (Oxford comma, em dash usage, etc.)
* Common anti-AI writing constraints that you apply universally
* Any defaults you always want the Agent to start from

### Project rules

Project rules live inside a specific project and only apply when you're working on it. Use them for:

* This project's POV and narrative distance
* Tense (past / present) and person (first / third)
* Project-specific stylistic goals or genre conventions
* Writing constraints specific to this story's tone

## Where rules live

```text theme={null}
your-novel/
├── .soloent/
│   └── rules/                 # Project rules (this project only)
│       ├── pov.md             # POV, tense, narrative distance
│       └── style.md           # Project-specific style requirements
├── SOLOENT.md                 # Story constitution: world, characters, plot, state
├── chapters/
└── ...
```

Global rules are stored in SoloEnt's system-level Rules directory and are loaded automatically for every project.

SoloEnt processes all `.md` files inside `.soloent/rules/`, combining them with any active global rules into a unified set of instructions for the Agent.

## Creating rules

<img alt="Rules panel showing Global Rules and Workspace Rules sections with a new rule file input and toggle controls" />

<Steps>
  <Step title="Open the Rules panel">
    Click the fourth icon from the left in the Agent panel，and switch the Rules manager.
  </Step>

  <Step title="Choose global or project">
    Select where the rule should live — globally (all projects) or just in the current project.
  </Step>

  <Step title="Create a new rule file">
    Click "New rule file...", enter a filename (e.g., `my-style`) and click on "+". The file will be created with a `.md` extension.
  </Step>

  <Step title="Write your rule">
    Add your instructions in markdown format. Keep each rule file focused on a single concern.
  </Step>
</Steps>

## Rules vs. SOLOENT.md

Rules are designed for **your habits and preferences** — the things that stay true regardless of which story you're working on. Story-specific content like world-building details, character profiles, and plot structure lives in [`SOLOENT.md`](/tips/SoloEnt), where the Agent can manage and update them as your project evolves.

These two systems are complementary:

|                      | Rules                                              | SOLOENT.md                                    |
| -------------------- | -------------------------------------------------- | --------------------------------------------- |
| **What it captures** | Your writing habits & preferences                  | Your story's content & state                  |
| **Scope**            | You as a writer                                    | This specific project                         |
| **Examples**         | Sentence rhythm, anti-AI constraints, POV defaults | World logic, character profiles, plot outline |
| **Updated by**       | You manually                                       | Mostly auto-updated by the Agent              |
| **Applies to**       | All projects (global) or this project              | This project only                             |

<Tip>
  Think of it this way: SOLOENT.md is the story's brain. Rules are your brain — your instincts, your style, your habits and non-negotiables as a writer.
</Tip>

### Toggling rules

<img alt="Rules panel showing an active rule file with the toggle switch highlighted" />

Every rule has a toggle to enable or disable it without deleting the file. This is useful when a specific task needs the Agent to temporarily step outside your usual constraints — for example, disabling a "short sentence" rule when writing a stream-of-consciousness passage.

## What to put in Rules

### Personal style habits (global)

These are the writing instincts you've developed as a writer. They should feel like second nature to you, even if you've never written them down before.

**Sentence rhythm example:**

```markdown theme={null}
# Sentence Style

- Default to shorter sentences (under 20 words) in scenes with tension or action
- Allow longer sentences in memory or introspective passages, but break them if they exceed 40 words
- Never write three long sentences in a row without a short one to interrupt the rhythm
- Vary sentence openings — avoid starting consecutive sentences with "She" or "He"
```

**Punctuation habits example:**

```markdown theme={null}
# Punctuation Preferences

- Use the Oxford comma
- Avoid em dashes for parenthetical asides — use commas or restructure the sentence instead
- Ellipses only for trailing-off speech or thought; not for dramatic pauses
- Semicolons are allowed in narration but not in dialogue
```

### Anti-AI writing constraints (global)

AI models have recognizable default patterns that flatten prose and signal non-human authorship. Add the ones that bother you most as standing constraints.

```markdown theme={null}
# Anti-AI Writing Constraints

## Overused words to avoid
- tapestry, mosaic, testament, dance (as metaphor), embrace (as metaphor)
- "It was a reminder that..." / "It was a testament to..."
- whisper (unless literal), shimmer, gleam
- "[X] hung between them" / "a surge of [X] rose in her chest" / "her heart sank" (stock somatic descriptions)
- "His eyes / expression / tone said it all"

## Structural patterns to avoid
- Do not summarize emotions in a final sentence after showing them in a scene
- Do not end scenes with a character reflecting on what just happened ("She realized that...")
- Avoid the "X, Y, and Z" three-item list pattern in narration
- No sentences structured as "Not [X]. [Y]." unless used very sparingly
- No excessive countdown timestamps: "X days until the tournament" / "one year and three months in total"
- No countdown phrasing: "three days left" / "two days remaining"
- No single-word sentences used for dramatic effect; no parallel clauses made of single words

## Character voice
- Do not repeat the same word or sentence pattern for a character unless it's an intentional verbal tic
- Maintain distinct voices per character — clearly different in register, rhythm, and tone
- The protagonist is allowed to be sarcastic, self-deprecating, and irritable

## Chapter openings and endings
- Do not open a chapter with a time marker: "One day later", "Three days before the match"
- Do not open a chapter with "He/She felt..."
- Do not end a chapter with an extended internal monologue
- Do not end a chapter with thematic reflection or moral commentary
- Every chapter ending must have a hook that pulls the reader forward
- Do not soften the end of a tense scene with a light or hopeful beat
```

### Project POV and narrative basics (project)

Every project has a fixed set of foundational writing decisions that the Agent should never deviate from.

```markdown theme={null}
# Project POV & Narrative Basics

## Point of View
- Close third-person limited
- One POV character per chapter; never switch mid-chapter
- The narrator knows only what the POV character knows — no dramatic irony from outside their perspective

## Tense & Person
- Past tense throughout
- Third person throughout; no second person

## Narrative Distance
- Stay close; render interiority directly
- Avoid free indirect discourse that could be mistaken for omniscient narration
```

### Project-specific style requirements (project)

Use this for genre conventions, tonal targets, or stylistic goals that are particular to this story.

```markdown theme={null}
# Project Style — Psychological Horror

## Tone targets
- Dread through implication, not explicit description
- The wrongness should accumulate slowly; avoid sudden revelation until Act III
- Understated prose; the horror lives in what's not said

## What to avoid in this project
- No jump-scare structure (sudden loud event after quiet)
- No gore for its own sake; visceral detail only if it serves the character's psychological state
- Avoid explaining the supernatural — ambiguity is the primary tool
```

## Writing effective rules

* **Be specific, not aspirational.** "Write beautifully" tells the Agent nothing. "Prefer concrete sensory detail over abstract emotional labeling" is actionable.
* **State prohibitions directly.** The most useful rules are often negative: "never do X." The Agent responds well to clear constraints.
* **Include the why when it's not obvious.** "Avoid em dashes (they read as an AI writing signature)" helps the Agent understand the intent and generalize it correctly.
* **Keep rules short.** Rules consume context tokens. A rule file that runs to five pages will crowd out your actual writing. If a rule needs extensive explanation, it's probably better handled in SOLOENT.md or a dedicated reference document.
* **One concern per file.** Split rules by topic so you can toggle individual concerns on and off without affecting the rest.

<Warning>
  Don't use Rules to duplicate what's already in SOLOENT.md. If you've defined a character's voice in `SOLOENT.md`, there's no need to re-state it in a rule — the Agent reads both.
</Warning>

## Conditional rules

Conditional rules activate only when you're working with files that match a path pattern. This lets you apply different writing constraints at different stages of a project without having to manually toggle rules each time.

As your rule library grows, loading every rule for every request wastes context tokens. Conditional rules keep the Agent focused on only the instructions that matter for the files you are currently working with.

A common use case: applying sentence rhythm and anti-AI constraints only when writing chapter files, so they don't interfere with notes or outlines.

```yaml theme={null}
---
paths:
  - "chapters/**"
---

# Prose Style

## Sentence rhythm
- Vary sentence length deliberately — short sentences for impact, longer ones for accumulation
- Never open three consecutive sentences with the same subject
- Break any sentence over 35 words into two

## Patterns to avoid
- No em dashes for parenthetical asides; restructure the sentence instead
- Avoid "she realized / she noticed / she understood" — show the realization, don't name it
- No three-item lists in narration (X, Y, and Z)
- Do not soften the end of a tense scene with a reflective or hopeful beat
```

### How it works

Add YAML frontmatter at the top of any rule file. The Agent evaluates each rule's `paths` against the files you're currently working with, and activates matching rules automatically.

**Rules without frontmatter are always active** — the right choice for your universal style habits and anti-AI constraints.

## Troubleshooting

<AccordionGroup>
  <Accordion title="The Agent keeps writing in patterns I've told it to avoid">
    **Most likely cause**: The rule is either toggled off or the instruction isn't specific enough.

    * Check that the rule file is toggled on in the Rules panel
    * Rewrite the constraint as a direct prohibition: "Do not..." rather than "Try to avoid..."
    * If it's a global rule, verify it's not being overridden by a project rule with conflicting instructions
  </Accordion>

  <Accordion title="A conditional rule isn't activating">
    **Check the following:**

    * The file you're editing must match the `paths` glob pattern
    * The YAML frontmatter must have proper `---` delimiters on both sides
    * The rule must be toggled on in the Rules panel
  </Accordion>

  <Accordion title="Rules and SOLOENT.md seem to be giving the Agent conflicting instructions">
    **This usually means there's overlap** — you've defined something in both places.

    * Keep story content (who a character is, what the world's rules are) in SOLOENT.md
    * Keep your writing preferences (how prose should feel, what patterns to avoid) in Rules
    * If you have stylistic requirements specific to this story's tone, a project rule is fine — just make sure it doesn't contradict SOLOENT.md's Section 5 (Stylistic Guidelines)
  </Accordion>

  <Accordion title="Too many rules are slowing things down or causing the Agent to miss instructions">
    Rules consume context tokens. If you have many active rules:

    * Toggle off rules that don't apply to your current task
    * Consolidate overlapping rules into a single file
    * Move any story-specific content that's crept into Rules back into SOLOENT.md where it belongs
  </Accordion>
</AccordionGroup>

## More Advanced Tips

<CardGroup>
  <Card title="SOLOENT.md Guide" icon="book-sparkles" href="/en/tips/SoloEnt">
    Understand what belongs in SOLOENT.md vs. Rules
  </Card>

  <Card title="Commands" icon="terminal" href="/en/tips/commands">
    Use `/review` to check whether your latest chapter follows your active rules
  </Card>
</CardGroup>


# Skills
Source: https://docs.soloent.ai/en/tips/skills

Package specialized Agent knowledge into reusable instruction sets that activate only when you need them

Skills are markdown files that give the Agent deep, task-specific knowledge for a particular type of work. Define one for short story creation, one for analysis the book, one for screenplay formatting — and the Agent loads only the skill that's relevant to what you are doing right now.

## How skills work

Skills use progressive loading to keep context lean. When you open a project, the Agent sees a list of available skills with their names and descriptions — a few dozen words each. That is all it loads at startup.

When you ask for something that matches a skill's description, the Agent activates that skill and loads its full instructions. A skill for short story creation stays dormant while you are working on a query letter. A skill for horror fiction does not consume context when you are outlining a thriller.

| Level                | When loaded             | Content                                 |
| -------------------- | ----------------------- | --------------------------------------- |
| **Metadata**         | Always, at startup      | Name and description from frontmatter   |
| **Instructions**     | When skill is triggered | Full SKILL.md body                      |
| **Supporting files** | As needed               | Templates, scripts, reference documents |

## Where skills live

Every skill is a directory containing a `SKILL.md` file. The directory name becomes the skill name. Skills can also include supporting files in three optional subdirectories.

```text theme={null}
your-novel/
├── .soloent/
│   └── skills/                  # Workspace skills (this project only)
│       └── short-story/
│           ├── SKILL.md                   # Required: name, description, instructions
│           ├── templates/                 # Optional: output templates, format guides
│           │   ├── chapter-outline.md
│           │   └── character-profile.md
│           ├── docs/                      # Optional: reference material, lookup tables
│           │   └── genre-conventions.md
│           └── scripts/                   # Optional: processing scripts the Agent can run
│               └── word-count.py
├── SOLOENT.md
├── chapters/
└── ...
```

Supporting files are loaded only when the instructions reference them — they do not consume context at startup. A template stays dormant until the Agent reaches the step that calls for it. A reference document is read only when the Agent needs to look something up.

**Global skills** are stored in the system-level skills directory and are available across all your projects. Use them for general-purpose capabilities you apply consistently — a skill for generating opening hooks, or one that guides the Agent through your standard manuscript review process.

**Workspace skills** live in `.soloent/skills/` at your project root. Use them for capabilities tied to a specific project or genre — a skill built around the particular rules of your romance series, or one calibrated to the explicit level and narrative register of your current story.

<Warning>
  When a global and workspace skill share the same name, the workspace skill takes precedence.
</Warning>

### Toggling skills

Every skill has a toggle to enable or disable it without deleting the file. Disable a skill when you do not need it for the current task — this keeps it out of context and avoids unintended activation.

## Creating a skill

The `SKILL.md` file has two parts: a frontmatter block (metadata) and the instruction body.

```markdown theme={null}
---
name: my-skill
description: Describe when to trigger this skill, not what it does.
---

Instructions for the Agent to follow when this skill is activated.
```

Two required frontmatter fields:

* `name` must **exactly** match the directory name
* `description` describes **when to trigger** this skill, not what it does (under 1,024 characters)

<img alt="SKILL.md file open in the editor showing the frontmatter fields name and description, with the Skills panel on the right displaying the english-short-story skill listed under Workspace Skills" />

<Steps>
  <Step title="Open the Skills menu">
    Click the fourth icon from the left in the Agent panel，and switch the Skills manager.
  </Step>

  <Step title="Create a new skill">
    Click "New skill..." and enter a name (e.g., `short-story`) and click on "+" . The Agent creates a skill directory with a template `SKILL.md` file.
  </Step>

  <Step title="Write your skill instructions">
    Edit the `SKILL.md` file:

    * Update the `description` to specify when this skill should trigger
    * Add your instructions, checklists, and workflow steps in the body
    * Optionally add supporting files in `templates/` or `docs/` subdirectories
  </Step>
</Steps>

<Tip>
  This is the official `english-short-story` skill file — a **domain expert** type skill with a complete 6-step workflow, checklists, and stop conditions. Copy and paste it directly.
</Tip>

<Accordion title="Get full english-short-story skill — click to expand and copy">
  ```markdown theme={null}
  ---
  name: english-short-story
  description: A SoloEnt official writing skill. Create English short fiction with strong hooks, sharp reversals, and high emotional intensity using a 6-step workflow: concept brainstorm, character design, outline, chapter writing with auto-review, and optional chapter review and title/hook generation. Explicit scenes and dark emotional registers are available tools, not requirements. Use when writing short stories, drafting chapters, designing characters, building outlines, reviewing fiction chapters, or generating titles and hooks.
  ---

  # English Short Story

  Manages the full creation workflow for English short fiction (10–20 chapters, 900–1,200 words each). Designed for stories built on strong hooks, sharp reversals, and high emotional intensity — across any register including thriller, horror, romance, erotic, literary, or hybrid. Explicit scenes and dark emotional content are available tools calibrated per story, not default requirements.

  ---

  ## Before Any Task: Read SOLOENT.md

  If `SOLOENT.md` exists, read it and extract:

  - **Section 8.1 Todo List** — what is the current pending task?
  - **Section 7.1 Current Status** — which step/chapter are we on?
  - **Section 7.4 Pending Threads** — active hooks and next chapter goal
  - **Section 3 Character Index** — is the Character Voice Profile available?

  If `SOLOENT.md` does not exist, proceed directly to Step 1. After the user confirms the core concept, run `/init` to create and populate `SOLOENT.md` before Step 2 begins.

  ---

  ## Usage

  Use this skill whenever you are working on an English short fiction project that prioritizes strong hooks, emotional intensity, and sharp plot reversals. It manages the full creation process from a one-line premise through to finished chapter files. Works across all registers — thriller, horror, romance, erotic, literary, or hybrid. Explicit content level is calibrated per story during Step 1.

  **Start a new story**: provide a one-line premise and the skill walks you through Steps 1–4 in sequence, requiring your confirmation at each stage before advancing. Steps 5 and 6 are optional and run on request.

  Trigger phrases that activate this flow:
  - Direct story ideas: "A woman finds a letter in her dead husband's coat…", "Two strangers trapped in a lighthouse…"
  - Intent statements: "I want to write a short story", "I have a story idea", "Let's write a short fiction"
  - Explicit skill invocation: "Use the short story skill", "Start the writing workflow", "短篇技能"

  **Resume an existing story**: open a project that already has a `SOLOENT.md` and say "continue the story" or "pick up where we left off" — the skill reads current state and resumes from the correct step.

  **Jump to a specific step**: name the step directly (e.g. "run Step 4 for Chapter 3", "review Chapter 5") and the skill starts there.

  **Prompt templates are optional**: if `templates/Prompt-0X-*.md` files are present the skill loads them; if not, it runs on the built-in instructions and checklists.

  ---

  ## The 6-Step Workflow

  ### Step 1 — Core Concept Brainstorm

  **Input**: One-line premise from user
  **Output**: Core concept + story arc (3+ reversals + explicit calibration + sensory register)

  If `templates/Prompt-01-Core Concept Brainstorm.md` exists, load and follow it. Otherwise proceed using the output spec and checklist below.

  **Checklist before advancing to Step 2**:
  - [ ] Central tension defined (desire / fear / obsession / transgression)?
  - [ ] 3+ major plot reversals designed?
  - [ ] Erotic / horror / thriller elements are structural, not decorative?
  - [ ] Explicit level calibrated and noted?
  - [ ] First-person narration viable?

  **Wait for user confirmation.**

  → Once confirmed, run `/init` to initialize `SOLOENT.md` before proceeding to Step 2.

  ---

  ## /init — Initialize SOLOENT.md

  Run automatically after Step 1 is confirmed, before Step 2 begins.

  **Actions**:
  1. Create `SOLOENT.md` in the project root.
  2. Fill in the following fields from the Step 1 output:
     - **Section 1 Project DNA**: genre, core themes, emotional promise
     - **Section 5 Stylistic Guidelines**: base tone, language register, pacing notes derived from concept
     - **Section 7.1 Current Status**: story title, story time, immediate atmosphere
     - **Section 8.1 Todo List**: check off `[x] Confirm core concept (Step 1)`, leave remaining todos open
     - **Section 8.2 Milestone Log**: add `[Completed] Step 1 — core concept confirmed: [premise summary] / explicit level: [X]`
  3. Confirm to user: `SOLOENT.md initialized — ready for Step 2.`

  ---

  ### Step 2 — Character Design

  **Input**: Confirmed core concept
  **Output**: Main cast + Character Voice Profile

  If `templates/Prompt-02-Character Design.md` exists, load and follow it. Otherwise proceed using the output spec and checklist below.

  **Rules**:
  - Complete and confirm characters **before** touching the outline
  - Must output a **Character Voice Profile**: 1–2 sentences per character (speech patterns + signature phrases) — pasted into every Step 4 prompt

  **Checklist before advancing to Step 3**:
  - [ ] Protagonist is active? Clear goal driving them forward?
  - [ ] Every character embodies a value, desire, fear, or moral position?
  - [ ] Protagonist's voice is distinct, layered, and evolves?
  - [ ] Character Voice Profile output and ready for Step 4?
  - [ ] Core cast ≤ 4 characters?
  - [ ] Relationships generate natural conflict?

  **Wait for user confirmation.**

  → Update `SOLOENT.md`:
  - **Section 3.1 Character Index**: add each character entry with name, age, role, and pointer to `characters/[name].md`
  - **Section 3.2 Relationship Network**: fill in key relationships
  - **Section 7.2 Character Snapshot**: fill in current character states
  - **Section 8.1**: check off character design todo, add `[ ] Complete story outline`
  - **Section 8.2**: log `[Completed] Step 2 — characters confirmed: [names] / Voice Profile ready`

  ---

  ### Step 3 — Outline and Chapter Plan

  **Input**: Confirmed concept + confirmed characters + chapter count (default 10–20)
  **Output**: 3-act outline + chapter-by-chapter plan

  If `templates/Prompt-03-Outline and Chapter Plan.md` exists, load and follow it. Otherwise proceed using the output spec and checklist below.

  **Chapter plan format** (must match Step 4 input):
  - Story beats: 3–5 per chapter
  - Intensity peak: note trigger method (dialogue / action / discovery)
  - Closing hook: note tension type (reversal / revelation / threat / disappearance)
  - No pre-set timestamp transitions

  **Checklist before advancing to Step 4**:
  - [ ] Chapter count confirmed with user?
  - [ ] Chapter 1 opens on immediate conflict?
  - [ ] Chapter 1 intensity peak strong enough to lock the reader?
  - [ ] Every chapter has: beats + intensity peak (trigger noted) + closing hook (tension type noted)?
  - [ ] Reversal points clearly marked?
  - [ ] Explicit / horror / thriller peaks distributed throughout, not clustered?

  **Wait for user confirmation.**

  → Update `SOLOENT.md`:
  - **Section 4.1 Story Structure**: fill in chapter range and act breakdown
  - **Section 4.2 Outline Index**: add pointer `[Story Outline] -> outline/master-outline.md`
  - **Section 8.1**: check off outline todo, add `[ ] Begin writing Chapter 1`
  - **Section 8.2**: log `[Completed] Step 3 — outline confirmed: [X] chapters / [act structure summary]`

  ---

  ### Step 4 — Chapter Writing + Auto-Review

  **Input**: Current chapter plan + Character Voice Profile; from Chapter 2 onward, also paste previous 1–2 chapters
  **Output**: Chapter prose (900–1,200 words) + auto-review report

  If `templates/Prompt-04-Chapter Writing.md` exists, load and follow it. Otherwise proceed using the output spec and checklist below.

  **Per-chapter process**:
  1. Chapter 1: paste chapter plan + Voice Profile
  2. Chapter 2+: paste chapter plan + Voice Profile + previous 1–2 chapters
  3. Run `wc -w chapters/ch-XX.md` for word count — do not use model self-estimate
  4. Output auto-review immediately after prose

  **Auto-Review Checklist** (run after every chapter, mark ✅/❌):
  - [ ] Word count 900–1,200 (verified via `wc -w`)?
  - [ ] All beats executed in order? Intensity peak landed with correct trigger?
  - [ ] Closing hook sharp? No moral summary / reflection / epiphany ending?
  - [ ] First-person POV clean — no "he thought / she felt" violations?
  - [ ] No 3+ consecutive introspective sentences?
  - [ ] Dialogue drives plot (≥ 50%)?
  - [ ] Each character's speech matches their Voice Profile?
  - [ ] No repeated phrasing within the same character's lines?
  - [ ] No timestamp scene openers ("The next day…" / "Three weeks later…")?
  - [ ] Age/occupation/backstory communicated obliquely after first mention?
  - [ ] Explicit intensity consistent with Step 1 calibration?
  - [ ] No AI tells (mechanical structures, em-dash overuse, clichéd phrasing)?

  → After each chapter, update `SOLOENT.md`:
  - **Section 7.1**: update `Current Chapter` to next chapter number
  - **Section 7.3 Recent Plot Flow**: replace with 3–5 key events from this chapter
  - **Section 7.4 Pending Threads**: update `Urgent Hooks`, `Active Foreshadowing`, `Next Chapter Goal`
  - **Section 7.5 Execution Review**: paste auto-review checklist result + word count
  - **Section 8.1**: check off current chapter todo, add next chapter todo
  - **Section 8.2**: log `[Completed] Ch-[X] written — [one-line plot summary]`

  > For deep-dive review of specific chapters, use Step 5 (optional).

  ---

  ### Step 5 — Chapter Review (Optional)

  **Input**: User-specified chapter(s)
  **Output**: Detailed review report + revision recommendations

  If `templates/Prompt-05-Chapter Review.md` exists, load and follow it. Otherwise proceed using the checklist below.

  **When to trigger**:
  - User questions a chapter's quality
  - Reversal chapters need tonal polish
  - Multi-chapter continuity check needed
  - Explicit content consistency review across chapters

  **Checklist**:
  - [ ] Word count (900–1,200) + outline fidelity (beats + intensity peak)?
  - [ ] Narrative logic continuity, prose style consistency, character voice fidelity?
  - [ ] ⚠️ Tag repetition: age/occupation ≤ 2 mentions; timestamp openers = 0; no phrase repetition?
  - [ ] Explicit escalation correct? Any tonal inconsistencies?
  - [ ] No AI tells (mechanical structures, digressive monologues, em-dash overuse)?

  → Update `SOLOENT.md`:
  - **Section 7.5 Execution Review**: log review findings and revision status for reviewed chapter(s)
  - **Section 8.1**: check off review todo if applicable

  ---

  ### Step 6 — Title and Opening Hook (Optional)

  **Input**: Completed full story (or concept + key scene summary)
  **Output**: 100–150 word opening hook + 3 title options

  If `templates/Prompt-06-Title and Hook.md` exists, load and follow it. Otherwise proceed using the output spec and checklist below.

  **Run only when user requests.** Recommended after all chapters are complete — extract tension from the actual draft.

  **Checklist**:
  - [ ] Opening hook is 100–150 words, names core conflict, creates immediate suspense?
  - [ ] 3 titles differ in style (conflict angle / intrigue angle / atmosphere angle)?
  - [ ] Titles calibrated for English-language fiction platforms?

  **Delivery**: individual chapter files + title and hook; merge into single file only if user requests.

  → Update `SOLOENT.md`:
  - **Section 7.1**: set `Current Chapter` to `Complete`
  - **Section 8.1**: check off Step 6 todo if present
  - **Section 8.2**: log `[Completed] Step 6 — title and hook delivered / project complete`

  ---

  ## Core Creative Principles

  - **Hook before everything**: the opening must create immediate tension — a question, a threat, a desire the reader cannot ignore
  - **Emotional intensity over social melodrama**: drive the story with visceral feeling — desire, dread, obsession, grief, rage — not domestic drama for its own sake
  - **Escalating stakes**: each act pushes further than the reader expects — in intimacy, danger, revelation, or moral cost
  - **Active protagonist**: story engine is what the protagonist *chooses to do*, not what happens to them
  - **Characters embody value**: each character represents a desire, a fear, a moral position, a kind of power
  - **3+ major reversals** distributed across the full story
  - **First-person narration** for maximum immersion
  - **Explicit content is a calibrated tool**: level set in Step 1 — can range from zero to high depending on the story's needs
  ```
</Accordion>

## Three common skill types

### Type 1: Checklist type

Best for verifying multiple criteria at a fixed point — so nothing gets missed. For example, a pre-submission check after each chapter is done:

```markdown theme={null}
---
name: chapter-submission-check
description: Use before finalizing a chapter to verify word count, consistency, and all checklist items.
---

## Pre-submission check (all must pass before advancing)
- [ ] Word count 900–1,200 (verified with tool, not estimated)
- [ ] All chapter beats executed, intensity peak landed
- [ ] Closing hook sharp — no moral summary or reflection ending
- [ ] Character name consistency check passed
- [ ] SOLOENT.md updated

## Output
Pass / Fail per item. Any Fail must be fixed before proceeding.
```

### Type 2: Workflow type

Best for operations with a fixed sequence of steps and a confirmation step before execution.

```markdown theme={null}
---
name: chapter-rewrite
description: Rewrite a chapter from scratch. Run only when explicitly requested.
---

## Steps
1. Backup: save the current chapter as ch-XX-original.md to preserve the original
2. Clarify: summarize the core direction of this rewrite in one paragraph and wait for user confirmation before writing
3. Execute: write the full chapter in the new direction (900–1,200 words)
4. Verify: run the chapter checklist against the original and output the main differences between old and new
```

### Type 3: Domain expert type

Best for situations where you want the Agent to work from expert experience rather than guessing. For example, a diagnosis skill for when the writing gets stuck:

```markdown theme={null}
---
name: story-diagnosis
description: Use when a chapter feels flat, pacing is off, or a scene isn't landing.
---

## Evidence collection
1. Read the full text of the problem chapter
2. Cross-check against the chapter beat plan in SOLOENT.md Section 4
3. Check the emotional arc of the chapters immediately before and after

## Decision matrix
| Symptom | Check first |
|---|---|
| Pacing drags | Beat density / any dead paragraphs |
| Emotional flatness | Character voice consistency / too much internal monologue |
| Closing hook fails | Tension type / has tension already been released within the chapter |

## Output format
Root cause / blast radius / fix steps / verification method
```

## Full example: Book summary

**This is a domain expert skill.** It defines a content-retrieval path (a script extracts chapters on demand), specialized knowledge (chapter recognition patterns in `docs/`), and a fixed output format (a note template in `templates/`) — the three elements that characterize a domain expert skill.

This skill extracts structured reading notes from a local book file — chapter summaries, key insights, and quotations — and saves them as formatted Markdown.

### Directory structure

```text theme={null}
book-summary/
├── SKILL.md                    # Workflow instructions
├── templates/
│   └── note-template.md        # Output format for the reading notes
├── docs/
│   └── chapter-patterns.md     # Chapter heading patterns for Chinese and English books
└── scripts/
    └── analyze-book.py         # Preprocesses book files before the Agent reads them
```

### Why this skill needs a script

Books are long. Loading an entire novel into context to extract its chapter structure would be wasteful and slow. The preprocessing script solves this:

* **`info` command** — reads the first 50 lines to identify the title, encoding, and total line count
* **`chapters` command** — scans the file for chapter headings and returns their line numbers, without reading the full text
* **`extract` command** — reads only the lines for a specific chapter, so the Agent processes one chapter at a time

The Agent calls the script at each stage of the workflow. Only the script's output enters context — not the book itself.

### Why this skill needs a docs file

Chapter heading formats vary widely across Chinese and English books. Rather than embedding all the recognition patterns in `SKILL.md`, the skill stores them in `docs/chapter-patterns.md`. The Agent reads this reference file when it needs to identify chapter boundaries — and ignores it otherwise.

Keeping reference material in `docs/` has two benefits: it keeps `SKILL.md` focused on workflow logic, and it lets you update the patterns without touching the main instructions.

### Why this skill needs a template

The output format is fixed: every set of reading notes follows the same structure — book overview, chapter summaries with key points and quotations, book-wide highlights, and a personal reflection section. Rather than re-specifying this format in every prompt, the skill stores it in `templates/note-template.md`.

When Step 5 (final assembly) runs, the Agent loads the template and fills in the placeholders. The format is consistent every time, and you only need to maintain it in one place.

### How the SKILL.md references supporting files

The instructions in `SKILL.md` point to each supporting file explicitly:

```markdown theme={null}
### Step 1: Analyze book structure

Run the preprocessing script to get book information:
  python3 scripts/analyze-book.py <file-path> info

### Step 5: Assemble notes

Output using the template: templates/note-template.md

For chapter heading recognition patterns, refer to: docs/chapter-patterns.md
```

## Writing effective skills

### Write "when to use me", not "what I do"

These sound similar, but they have a significant impact on trigger accuracy.

```yaml theme={null}
# Wrong — describes capability, not trigger condition
description: This skill helps improve writing quality and checks chapters for common issues.

# Right — describes when to activate
description: Use when reviewing a completed chapter draft for pacing, voice, and consistency issues.
```

Descriptions **stay in context at all times** — every enabled skill is continuously consuming your token budget. Keep them short, keep them specific, and state the trigger scenario directly:

```yaml theme={null}
# Inefficient (~45 tokens)
description: |
  This skill helps you review and improve your writing chapters.
  It checks for common issues like pacing problems, character voice...
  Use this when you want to ensure your chapter quality before proceeding.

# Efficient (~9 tokens)
description: Use when reviewing a chapter for quality issues.
```

### Complete steps, inputs, outputs, and stop conditions

A skill instruction that only writes the beginning of a process — with no clear completion signal — leaves the Agent to decide when to stop. That is rarely where you want it to stop. Every step should have:

* **Input**: what context or files this step needs
* **Output**: what this step produces
* **Stop condition**: when to wait for confirmation, when to continue

### Keep the body lean — move bulk material to supporting files

Keep `SKILL.md` under 5,000 words. If you need large reference material — style guides, recognition patterns, character profiles — put them in the `docs/` subdirectory and reference them by filename. The Agent loads them only when needed, keeping them out of context otherwise.

### Decide auto-trigger based on call frequency

| Frequency                           | Recommended strategy                                                       |
| ----------------------------------- | -------------------------------------------------------------------------- |
| High (more than once per session)   | Keep auto-trigger on; refine description for more precise activation       |
| Low (at most once per session)      | Disable auto-trigger; invoke manually; description can stay out of context |
| Very low (less than once per month) | Remove the skill; document it in SOLOENT.md instead                        |

### Common pitfalls

| Problem                         | Symptom                                                                      | Fix                                                                        |
| ------------------------------- | ---------------------------------------------------------------------------- | -------------------------------------------------------------------------- |
| **Description too broad**       | `description: help with writing` — any writing request triggers it           | Narrow to a specific scenario; include actual trigger phrases              |
| **Body too long**               | Hundreds of lines of instructions packed into `SKILL.md`                     | Move reference material to `docs/`; keep only the skeleton in instructions |
| **One skill doing five things** | Same skill covers drafting, outlining, reviewing, diagnosing, and formatting | Split them; one domain per skill, each toggleable independently            |

## Troubleshooting

<AccordionGroup>
  <Accordion title="A skill isn't activating when I expect it to">
    **Most likely cause**: The description isn't specific enough, or the skill is toggled off.

    * Check that the skill is toggled on in the Skills panel
    * Rewrite the `description` to include the exact phrases and task types you use — "helps with writing" rarely triggers correctly; name the workflow, the output, and sample invocation phrases
    * Make sure the skill directory contains a valid `SKILL.md` with both `name` and `description` frontmatter fields
    * Verify that `name` in the frontmatter **exactly matches** the directory name
  </Accordion>

  <Accordion title="A skill keeps activating when I don't want it to">
    **The description is too broad** — it matches requests you didn't intend it for.

    * Narrow the `description` to describe only the specific tasks this skill handles
    * Toggle the skill off when you're working outside its domain
    * If you have multiple skills with overlapping descriptions, make each one more specific about when it should and shouldn't activate
  </Accordion>

  <Accordion title="The skill's instructions are being ignored or only partially followed">
    **Check for these common issues:**

    * The instructions may exceed the context limit — keep `SKILL.md` under 5,000 words and move reference material into `docs/` subdirectories
    * A conflicting global or project rule may be overriding part of the skill's behavior — check your active rules for contradictions
    * If the skill uses checklists, ensure each checklist item has an explicit "wait for confirmation" gate; without it the Agent may advance automatically
    * Supporting files referenced in the instructions must exist at the specified paths — a missing file causes that step to be skipped silently
  </Accordion>

  <Accordion title="Supporting files (templates, docs, scripts) aren't being loaded">
    **The Agent only loads supporting files when the instructions explicitly reference them.**

    * Check that the file path in the instruction matches the actual location (e.g., `templates/chapter-outline.md`, not just `chapter-outline.md`)
    * Verify the file exists inside the skill directory under the correct subdirectory (`templates/`, `docs/`, or `scripts/`)
    * If the file is never referenced by name in `SKILL.md`, the Agent has no reason to load it
  </Accordion>

  <Accordion title="A workspace skill and a global skill with the same name are conflicting">
    **Workspace skills always take precedence over global skills of the same name.**

    * This is expected behavior — if a workspace skill is active, the global skill with the same name is ignored entirely
    * If you want the global skill to apply, rename or remove the workspace skill
    * If both should coexist, give them distinct names and update their descriptions accordingly
  </Accordion>
</AccordionGroup>

## More advanced tips

<CardGroup>
  <Card title="Workflows" icon="diagram-project" href="/en/tips/workflows">
    Automate repeatable multi-step processes with workflow files
  </Card>

  <Card title="Rules" icon="scroll" href="/en/tips/rules">
    Capture your writing preferences so the Agent always writes in your voice
  </Card>

  <Card title="Rules vs. Workflows vs. Skills" icon="scale-balanced" href="/en/tips/compare">
    Three ways to instruct your Agent — learn the differences and when to use each
  </Card>
</CardGroup>


# Subagents
Source: https://docs.soloent.ai/en/tips/subagent

Let the main Agent automatically dispatch read-only subagents to explore in parallel — saving context, speeding up loading, and controlling cost

SoloEnt supports **collaboration between a main Agent and subagents**. Based on the task, the main Agent automatically dispatches subagents and assigns exploration work; multiple subagents run in **parallel** and then report their results back to the main Agent. This keeps the main Agent's context from being filled with large amounts of raw material during wide-ranging exploration.

## When it fits

<Check>
  **Good for**: knowledge exploration, information gathering, and classification/collaboration — the main Agent assigns different exploration tasks to different subagents to complete in parallel, then aggregates the results.
</Check>

* Gathering weapon references, worldbuilding knowledge, and character design separately
* Reading different chapters separately and then summarizing

<Warning>
  **Not good for**:

  * Assigning **repetitive** tasks to multiple subagents at once — different subagents repeatedly read the same context, which is wasteful
  * **Production/output** tasks — subagents focus on exploring, gathering, and organizing; they **have no write tools** and are not responsible for output
</Warning>

## Where to enable

In **Agent Settings** you can turn subagents on or off. When enabled, the main Agent will automatically trigger subagent collaboration as the task requires.

<img alt="The subagent toggle in Agent feature settings, described as letting SoloEnt run read-only research subagents in parallel to gather evidence from your writing project" />

For example, during wide-ranging knowledge exploration, the main Agent can delegate **read-only subagents** to run in parallel and aggregate the results back, greatly saving context.

<img alt="The main Agent running 3 subagents in parallel to analyze the previous 3 chapters, the characters, and the outline in support of writing the next chapter" />

## Configure a separate model for subagents

Subagents mainly explore the information the main task needs, so you can pair them with a **different model from the main Agent**. Benefits of tuning this well:

* Saves the main Agent's context
* Parallel exploration speeds up information loading
* Using a cheaper model lowers overall cost

<img alt="The SUBAGENT section in Agent Settings > API Configuration, where the subagent model can be set to Same as main or specified separately" />

<CardGroup>
  <Card title="Context Window" icon="gauge" href="/en/faq">
    Learn how context is consumed and when to open a new window
  </Card>

  <Card title="Skills" icon="wand-magic-sparkles" href="/en/tips/skills">
    Turn reusable exploration/organization flows into Skills to use with subagents
  </Card>
</CardGroup>


# Voice Input
Source: https://docs.soloent.ai/en/tips/voice-input

Speak directly in the chat window to make writing easier

<Note>
  Voice input is supported from **v0.12.0**.
</Note>

Click **Voice input** (the microphone icon) in the chat window and you can start speaking your input, which makes writing easier.

<img alt="The Voice input microphone button to the right of the chat input box" />

Click it to start speaking, then click again when you're done. Transcription into text starts automatically, and the input box shows **Transcribing...** while it works.

<img alt="The input box showing Transcribing... while speech is being transcribed" />

## Credit usage

Tokens consumed by Voice input are deducted automatically from your **membership quota**. You can review them under **usage details** in the Dashboard.


# Workflows
Source: https://docs.soloent.ai/en/tips/workflows

Package repetitive writing tasks into reusable step files and run an entire process with a single command

Workflows are Markdown files that define a series of steps to guide the Agent through complex or repetitive tasks. Type `/` followed by the workflow filename in the editor to invoke one — for example, `/bookcover.md`.

Writing is a multi-stage process, and each stage can be broken down into smaller steps to achieve better results. Drafting an outline, expanding it into a volume plan, then refining each batch of chapters through four or five quality passes — each of these tasks involves remembering a dozen steps, working in the right order, and updating files one by one. Miss a step and you start over. Workflows turn these multi-step processes into a single command. Type `/chapteroutline.md` and the Agent loads your story settings, produces chapter outlines, and waits for your confirmation before continuing. Every subsequent batch reuses the same command. You just review and decide.

## Invoking a workflow

Type `/` in the chat input and SoloEnt shows all available workflows. Keep typing to filter by filename — `/book` will match `bookcover.md`. Select a workflow and press Enter to start it.

<img alt="Typing /bo in the chat input, with autocomplete showing the matching bookcover.md workflow" />

The Agent executes each step in sequence, pausing when it needs a decision from you. You can stop a workflow at any point in the conversation.

<Tip>
  After completing a new task, tell the Agent: "Turn the process we just went through into a workflow file." The Agent analyzes the conversation, extracts the steps, and generates the workflow file. Experience you've built up becomes reusable automation.
</Tip>

## Where workflows live

Workflows can be stored in two locations: your writing project directory or a system-level global directory.

**Global workflows** are stored in the system-level directory and are available across all projects. Use them for general-purpose processes like cover prompt generation, standard manuscript review checklists, or publishing prep.

**Workspace workflows** live in `.soloent/workflows/` at your project root. Use them for processes tied to a specific project — the pacing logic for this book's chapter outlines, or the quality checks specific to this series.

```text theme={null}
your-novel/
├── .soloent/
│   └── workflows/         # Workspace workflows (this project only)
│       └── chapteroutline.md
├── SOLOENT.md
├── chapters/
└── ...
```

<Warning>
  When a global and workspace workflow share the same name, the workspace workflow takes precedence.
</Warning>

### Toggling workflows

Every workflow has a toggle to control whether it appears in the `/` menu, without deleting the file.

Workflows never run on their own — they only execute when you invoke them with a command.

## Creating a workflow

<img alt="Workflows panel showing a list of global workflows (bookcover.md, chapteroutline.md, outline.md) with toggle, edit, and delete controls for each" />

<Steps>
  <Step title="Open the Workflows menu">
    Click the fourth icon from the left in the Agent panel，and switch the Workflows manager.
  </Step>

  <Step title="Create a new workflow file">
    Click "New workflow file...", enter a filename (e.g., `bookcover`) and click on "+". The file is created with a `.md` extension.
  </Step>

  <Step title="Write your workflow steps">
    Add a title and steps in Markdown format. Describe what each step should accomplish.
  </Step>
</Steps>

## Workflow structure

A workflow is a Markdown file with a title and numbered steps. The filename is the command: `bookcover.md` is invoked with `/bookcover.md`.

Steps can be written at different levels of detail:

* **High-level instructions**: "Analyze the novel outline and extract the protagonist's character traits" — let the Agent decide how to execute
* **Precise control**: Use specific format requirements or templates — when you need a fixed output

## What workflows can contain

### Natural language instructions

Write steps in plain language and the Agent will interpret and execute them:

```markdown theme={null}
## Step 1: Gather novel basics
Read SOLOENT.md and extract: book title, key characters, world tone, and genre.
If the information is incomplete, ask the author and wait for their response before continuing.

## Step 2: Confirm cover direction
Based on the story's tone, propose three cover design directions and wait for the author to choose one.
```

This approach works well when the logic is clear but the execution details need the Agent to adapt.

### Precise template control

When you need a fixed output format, embed the template directly in the workflow:

````markdown theme={null}
## Step 3: Generate cover prompt

Use the following template to generate the prompt, replacing all `***` placeholders:

```
Generate a book cover for my novel "***".
Dimensions: 600×800 px (use 1800×2400 for print quality)

Design elements:
- Background:
- Color palette:
- Imagery:

Typography:
- Title: Bold, elegant serif font
- Author name: Clean, minimal sans-serif

Layout:
[Top 20%] Negative space with atmosphere
[Middle 40%] Title
[Bottom 30%] Author name
[Bottom 10%] Optional tagline in small text
```
````

## Example workflows

### Book cover prompt generation

This workflow reads the novel first before generating a cover prompt, so the result reflects the story's actual tone rather than a generic description.

```markdown theme={null}
# Book cover prompt

Generate a cover prompt for this novel.

## Step 1: Understand the novel
Read SOLOENT.md to understand the outline and main characters. If SOLOENT.md is missing or doesn't have complete story details, ask the author for:
- Book title and author name
- Genre and tone (bright / dark / fantasy / realist…)
- Key imagery or a defining scene

## Step 2: Generate the cover prompt

Use the following template, replacing all `***` with specific content:

Generate a book cover for my novel "***".
Dimensions: 600×800 px (use 1800×2400 for print quality)

Design elements:
- Background:
- Color palette:
- Imagery:

Typography:
- Title: Bold, elegant serif font with a light distressed texture
- Author name: Clean, minimal sans-serif
- Optional tagline: ""

Layout:
[Top 20%] Negative space with an aurora effect
[Middle 40%] Title
[Bottom 30%] Author name
[Bottom 10%] Optional tagline in small text
```

### Batch chapter outline generation

This workflow breaks a novel's chapter outlines into manageable batches. Each batch is confirmed before moving on, so you're never handed more than you can review at once.

````markdown theme={null}
# Batch chapter outline

Generate detailed outlines for 5 chapters at a time, confirming each batch before continuing.

**Target word count per chapter: 2,000–3,000 words**

## Step 1: Propose this batch

Read the story outline in SOLOENT.md and propose the chapter range for this batch (Chapter X through Chapter Y):
- State the chapter numbers this batch covers
- Reference the pacing from the story's structure (e.g., two chapters of tension before every major beat)
- Provide a one-sentence summary for each chapter

**Wait for the author to confirm the summaries before continuing.**

## Step 2: Expand each chapter outline

For each chapter in this batch, produce a full outline using the following format:

```markdown
## Chapter X: [Title]

**Summary**: [One sentence]

**Scene breakdown**:
- Scene 1: [What happens] + [Writing guidance: emotional tone, detail requirements]
- Scene 2: [What happens] + [Writing guidance: emotional tone, detail requirements]

**Foreshadowing & threads**:
- Plant: [New foreshadowing] (expected payoff in Chapter Y)
- Advance: [Existing thread being developed]
- Resolve: [Foreshadowing being paid off] (callback to Chapter Z)

**Ending hook**: [The cliffhanger or question that pulls the reader forward]
```

**Wait for the author to confirm this batch before continuing.**

## Step 3: Review settings and adjust

After completing a batch:

- **Power system check**: Can the existing mechanics support the major payoffs in this volume? If not, propose adjustments.
- **Character check**: Does this volume need new supporting characters? Does the protagonist show meaningful growth?

**Wait for the author's confirmation, then return to Step 1 for the next batch or conclude this volume's outlines.**
````

## Writing effective workflows

* **Start simple.** Write steps in natural language first. Only add precise templates when you need a fixed output format.
* **Be explicit about decision points.** If a step requires a choice from you, state it clearly: "Wait for the author to confirm before continuing."
* **Specify failure handling.** Tell the Agent what to do when something is missing: "If the information is incomplete, ask the author — do not assume."
* **Keep workflows focused.** `bookcover.md` handles covers. `chapteroutline.md` handles chapter outlines. Split complex processes into separate workflows that can run independently.

<Warning>
  Workflows execute with your permissions and can read and modify your project files. Review a workflow's contents before running it, especially if it comes from an external source.
</Warning>

## Workflows vs. Commands

The built-in commands (`/review`, `/market`) are workflows SoloEnt ships out of the box, designed by writing specialists and ready to use without any setup. Workflows are ones you define yourself, fully customized to your project and process.

The two complement each other: use built-in commands for standardized quality checks, and custom workflows to automate what's unique to your writing process.

## Troubleshooting

<AccordionGroup>
  <Accordion title="My workflow doesn't appear in the / menu">
    **Check the following:**

    * The workflow must be toggled on in the Workflows panel — toggled-off workflows are hidden from the menu
    * The file must have a `.md` extension and be located in either the global workflows directory or your project's `.soloent/workflows/` folder
    * If you just created the file, try closing and reopening the panel to refresh the list
  </Accordion>

  <Accordion title="The Agent is skipping steps or not pausing for confirmation">
    **The Agent advances automatically unless explicitly told to wait.**

    * Add a clear pause instruction at every decision point: "Wait for the author to confirm before continuing"
    * Avoid writing steps as a continuous block — break them into distinct numbered sections so the Agent treats each as a separate stage
    * If a step is being skipped entirely, check whether its instructions are reachable (not buried inside a template block the Agent misread as an example)
  </Accordion>

  <Accordion title="The workflow can't find my project files">
    **Most likely cause**: The file paths in the workflow don't match your actual project structure.

    * Use relative paths from the project root (e.g., `SOLOENT.md`, `chapters/ch-01.md`)
    * If the workflow reads `SOLOENT.md` and the file doesn't exist yet, the Agent may fail silently — add a fallback instruction: "If SOLOENT.md is missing, ask the author for the required information"
    * Workspace workflows run in the context of the open project folder — make sure the correct project is open when you invoke the command
  </Accordion>

  <Accordion title="A workspace workflow isn't overriding the global workflow of the same name">
    **Workspace workflows take precedence — but only if the names match exactly.**

    * Check that the filename (including extension) is identical: `chapteroutline.md` vs `chapteroutline.md`
    * A single character difference in the name means both workflows exist independently and neither overrides the other
    * Verify the workspace workflow is saved in `.soloent/workflows/` at the project root, not in a subdirectory
  </Accordion>

  <Accordion title="The workflow is modifying files I didn't expect it to touch">
    **Workflows run with your full permissions and can read and modify any file in your project.**

    * Review the workflow's steps before running it, especially if it came from an external source
    * If an unexpected file was modified, check whether any step contains a broad instruction like "update all chapter files" — make the scope explicit
    * For workflows that write output files, specify the exact target path in the step to prevent the Agent from choosing its own location
  </Accordion>
</AccordionGroup>

## More Advanced Tips

<CardGroup>
  <Card title="Commands" icon="terminal" href="/en/tips/commands">
    Learn about SoloEnt's built-in /review and /market commands
  </Card>

  <Card title="Rules" icon="scroll" href="/en/tips/rules">
    Capture your writing habits in Rules to use alongside your workflows
  </Card>
</CardGroup>


# API Configuration Issues
Source: https://docs.soloent.ai/en/troubleshooting/api-configuration

Troubleshoot common API configuration problems

Fix common issues when setting up API keys in SoloEnt.

## Invalid API key

**Problem:** Error message says "Invalid API key" or authentication fails.

**Solutions:**

* Verify you copied the entire key without spaces
* Check if key has expired or been revoked
* Regenerate a new key from provider dashboard
* Ensure you're using the correct key format for the provider

## Connection fails

**Problem:** Cannot connect to AI model or requests timeout.

**Solutions:**

* Check your internet connection
* Verify Base URL format (for OpenAI-compatible APIs)
* Try without VPN or proxy
* Check if the provider service is operational
* Verify firewall isn't blocking the connection

## Rate limit exceeded

**Problem:** Error message says "Rate limit exceeded" or "Too many requests".

**Solutions:**

* Wait a few minutes before trying again
* Upgrade your API plan for higher limits
* Switch to a different model temporarily
* Check your provider dashboard for usage limits

## Model not available

**Problem:** Selected model doesn't work or returns errors.

**Solutions:**

* Verify the model ID is correct
* Check if you have access to this model
* Ensure your API tier includes this model
* Try a different model from the same provider

## Missing required fields

**Problem:** Cannot save configuration, fields are incomplete.

**Solutions:**

* For standard providers: Only API key is required
* For OpenAI-compatible: All three fields required (Base URL, API Key, Model ID)
* Double-check provider documentation for exact requirements
* Contact provider support if unsure

<Card title="Back to API Configuration" icon="key" href="../quickstart/api-configuration">
  Return to the main API setup guide
</Card>


# Environment Issues
Source: https://docs.soloent.ai/en/troubleshooting/environment

Troubleshoot system environment and configuration problems

Fix common issues related to system environment, shell, and dependencies.

## PowerShell not recognized

**Problem:** Error message "PowerShell is not recognized as an internal or external command".

**Solutions:**

* **Windows users:**
  * Add PowerShell to system PATH
  * Open Settings → System → About → Advanced system settings → Environment Variables
  * Add PowerShell path: `C:\Windows\System32\WindowsPowerShell\v1.0\`
  * Restart SoloEnt after changing PATH

* **Alternative:**
  * Use Command Prompt (cmd) instead
  * Change default terminal in Settings
  * Reinstall PowerShell if missing

## Command not found

**Problem:** Terminal shows "command not found" or similar errors.

**Solutions:**

* Verify the command/tool is installed on your system
* Check if command is in system PATH
* Restart terminal after installing new tools
* Use full path to executable if needed
* Install missing dependencies

## Python/Node not found

**Problem:** Python or Node.js commands don't work.

**Solutions:**

* Install Python from python.org or Node.js from nodejs.org
* Verify installation: `python --version` or `node --version`
* Add to PATH if installed but not recognized
* Restart SoloEnt after installation
* Check for version conflicts (multiple Python/Node versions)

## Permission denied errors

**Problem:** "Permission denied" or "Access denied" errors.

**Solutions:**

* **macOS/Linux:**
  * Use `sudo` if appropriate
  * Check file permissions with `ls -la`
  * Grant Full Disk Access in System Settings

* **Windows:**
  * Run SoloEnt as Administrator if needed
  * Check folder security settings
  * Disable User Account Control temporarily

## Path with spaces causes errors

**Problem:** Commands fail when file paths contain spaces.

**Solutions:**

* Wrap paths in quotes: `"C:\My Documents\project"`
* Use escape characters in terminal
* Avoid spaces in project folder names
* Use hyphens or underscores instead: `my-project` not `my project`

## Git not available

**Problem:** Git commands don't work or Git is not found.

**Solutions:**

* Install Git from git-scm.com
* Verify installation: `git --version`
* Add Git to system PATH
* Restart terminal/SoloEnt after installation
* Use GitHub Desktop as alternative

## Environment variables not set

**Problem:** Application can't find required environment variables.

**Solutions:**

* Set variables in system settings (not just terminal session)
* **Windows:** System Properties → Environment Variables
* **macOS/Linux:** Add to `~/.zshrc` or `~/.bashrc`
* Restart SoloEnt after setting variables
* Verify with `echo $VARIABLE_NAME` (Unix) or `echo %VARIABLE_NAME%` (Windows)

## Shell compatibility issues

**Problem:** Commands work in one shell but not another.

**Solutions:**

* Check which shell you're using: `echo $SHELL`
* Switch to compatible shell (bash, zsh, PowerShell)
* Configure default shell in SoloEnt settings
* Use cross-platform commands when possible

## System architecture mismatch

**Problem:** Tools or binaries incompatible with your system.

**Solutions:**

* Check system architecture: `uname -m` (Unix) or `systeminfo` (Windows)
* Download correct version (x64, ARM64, etc.)
* Use compatible alternatives
* Enable Rosetta 2 on Apple Silicon Macs if needed

## Antivirus blocking operations

**Problem:** Antivirus software interferes with SoloEnt operations.

**Solutions:**

* Add SoloEnt to antivirus whitelist/exceptions
* Temporarily disable antivirus to test
* Check antivirus logs for blocked actions
* Contact antivirus support for proper configuration

## Disk space issues

**Problem:** Operations fail due to insufficient disk space.

**Solutions:**

* Check available disk space
* Clear temporary files and caches
* Move project to drive with more space
* Delete unnecessary files

## Network/firewall issues

**Problem:** Cannot connect to API services or download updates.

**Solutions:**

* Check firewall settings
* Allow SoloEnt through firewall
* Configure proxy settings if behind corporate firewall
* Test connection with different network
* Contact IT department for enterprise environments

<Tip>
  Many environment issues are resolved by restarting SoloEnt after making system changes like PATH updates or installing new tools.
</Tip>

## Still having issues?

<CardGroup>
  <Card title="Email Support" icon="envelope" href="mailto:team@soloent.ai">
    Contact our support team with environment details
  </Card>

  <Card title="Join Discord" icon="discord" href="https://discord.com/invite/NUh3skEFuY">
    Get help from the community
  </Card>

  <Card title="Quick Start" icon="rocket" href="../quickstart/quickstart">
    Return to setup guide
  </Card>
</CardGroup>

## Reporting environment issues

When reporting environment problems, include:

* **Operating system**: Windows 10/11, macOS version, Linux distribution
* **System architecture**: x64, ARM64, Apple Silicon
* **Shell type**: PowerShell, CMD, bash, zsh
* **Error messages**: Full text of any errors
* **Steps taken**: What you tried before reporting

This information helps us diagnose and resolve issues faster.


# File Conversion Issues
Source: https://docs.soloent.ai/en/troubleshooting/file-format-conversion

Troubleshoot document format conversion problems

Fix common issues when converting Word documents to Markdown.

## Conversion fails

**Problem:** Word document won't convert or shows error.

**Solutions:**

* Ensure file is valid `.docx` format (not `.doc` or other formats)
* Check file isn't corrupted or password-protected
* Try opening and resaving in Microsoft Word
* Remove password protection before conversion
* Convert to `.txt` first as a workaround
* Check file size (very large files may fail)

## Formatting looks wrong

**Problem:** Converted Markdown has formatting issues.

**Solutions:**

* This is normal for complex Word documents with advanced formatting
* Manually adjust Markdown formatting using the editor
* Simplify original Word document (remove complex styles, tables, etc.)
* Reconvert after simplification
* Use plain text paste for heavily formatted content

## Can't find converted file

**Problem:** Conversion completes but can't locate the `.md` file.

**Solutions:**

* Check the **same folder** as the original `.docx` file
* Look for file with same name but `.md` extension
* Refresh the file manager (reload icon)
* Search for `.md` files in project
* Try converting again

## Images missing after conversion

**Problem:** Images from Word document don't appear in Markdown.

**Solutions:**

* Images need manual placement in Markdown
* Extract images from Word document separately
* Place images in `images/` folder
* Add image references manually in Markdown: `![alt text](path)`

## Batch conversion not working

**Problem:** Cannot convert multiple files at once.

**Solutions:**

* Ensure **Markdown Hub** plugin is installed
* Verify all files in folder are `.docx` format
* Try converting files individually first
* Check folder permissions
* Restart SoloEnt and try again

## Special characters garbled

**Problem:** Special characters or non-English text appears incorrect.

**Solutions:**

* Check original Word document encoding
* Ensure UTF-8 encoding is used
* Resave Word document with correct encoding
* Manually fix characters in Markdown editor

<Card title="Back to File Conversion Guide" icon="file-import" href="../tips/file-format-conversion">
  Return to the main conversion guide
</Card>


# Login Issues
Source: https://docs.soloent.ai/en/troubleshooting/login

Troubleshoot authentication and sign-in problems

Fix common issues when signing in to SoloEnt.

## Browser doesn't open

**Problem:** Login page doesn't appear when launching SoloEnt.

**Solutions:**

* Check if browser window opened in background (check taskbar/dock)
* Look for login URL in SoloEnt window and copy manually
* Set your default browser if none configured
* Restart SoloEnt and try again

## Authentication fails

**Problem:** Cannot complete login process or credentials rejected.

**Solutions:**

* Verify your Google or Microsoft account credentials
* Try the other login method (switch between Google and Microsoft)
* Check your internet connection
* Disable VPN temporarily during login
* Clear browser cache and cookies
* Try a different browser

## Can't return to app

**Problem:** After successful login, cannot get back to SoloEnt.

**Solutions:**

* Ensure SoloEnt is still running in background
* Allow browser to open external applications in browser settings
* Manually switch to SoloEnt from taskbar/dock
* Restart SoloEnt and complete login again

## Login page loops

**Problem:** Keeps redirecting to login page repeatedly.

**Solutions:**

* Clear browser cookies for soloent.ai
* Try incognito/private browsing mode
* Check system date and time are correct
* Contact support if issue persists

<Tip>
  **For users in China:** Microsoft Account typically has better connectivity than Google Account due to network conditions.
</Tip>

<Card title="Back to Login Guide" icon="right-to-bracket" href="../quickstart/login">
  Return to the main login instructions
</Card>