> ## Documentation Index
> Fetch the complete documentation index at: https://docs.compasslabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Coding Assistant

> Point Cursor, Claude Code, Copilot, or any AI coding assistant at Compass's LLM-optimized docs and let it build the integration.

## Build with AI assistants

Compass provides LLM-optimized documentation that AI coding assistants can fetch directly. These endpoints follow the [llms.txt standard](https://llmstxt.org/) and are designed for AI consumption — concise, example-rich, and including all SDK field names and parameters.

<Info>
  Most AI coding assistants (Cursor, Claude Code, GitHub Copilot, ChatGPT) can automatically fetch and use these documentation URLs when you reference them in your project.
</Info>

***

## Documentation URLs

Point your AI assistant to these URLs:

| URL                                                                                | Description                                   | Size   |
| ---------------------------------------------------------------------------------- | --------------------------------------------- | ------ |
| **[docs.compasslabs.ai/llms-full.txt](https://docs.compasslabs.ai/llms-full.txt)** | Complete documentation with all examples      | \~60KB |
| **[docs.compasslabs.ai/llms.txt](https://docs.compasslabs.ai/llms.txt)**           | Condensed version for smaller context windows | \~17KB |

***

## Setup for AI coding assistants

<Tabs>
  <Tab title="Cursor">
    Add to your project's `.cursor/rules` or `.cursorrules` file:

    ```
    When working with Compass API, fetch documentation from:
    https://docs.compasslabs.ai/llms-full.txt
    ```
  </Tab>

  <Tab title="Claude Code">
    Add to your project's `CLAUDE.md` file:

    ```markdown theme={"system"}
    ## Compass API Documentation

    For Compass API integration, reference: https://docs.compasslabs.ai/llms-full.txt
    ```
  </Tab>

  <Tab title="Other Assistants">
    Reference the documentation URL in your prompt or project configuration:

    ```
    https://docs.compasslabs.ai/llms-full.txt
    ```
  </Tab>
</Tabs>

***

## What's included

The LLM documentation covers:

| Section                 | Description                                                       |
| ----------------------- | ----------------------------------------------------------------- |
| **SDK Setup**           | Installation for Python and TypeScript, environment configuration |
| **Core Concepts**       | Product Accounts, gas sponsorship, embedded fees                  |
| **API Endpoints**       | All V2 endpoints with parameters and responses                    |
| **SDK Field Reference** | Exact field names for both Python and TypeScript SDKs             |
| **Code Examples**       | Complete working examples for common workflows                    |
| **Troubleshooting**     | Common errors and solutions                                       |

***

## Example prompts

Once your AI assistant has access to the documentation, you can ask:

* "Write a Python script to fetch the top 5 USDC vaults on Base by APY"
* "How do I implement gas sponsorship for deposits?"
* "Create a TypeScript function to check all my earn positions"
* "What's the difference between `earn_transfer` and `earn_manage`?"

***

<CardGroup cols={3}>
  <Card title="CLI" icon="terminal" href="/v2/Agents/CLI">
    Prefer to drive Compass from the command line? The CLI exposes every endpoint as a subcommand.
  </Card>

  <Card title="Agent Skill" icon="puzzle-piece" href="/v2/Agents/Agent-Skill">
    Let an AI agent run the CLI for you — install the packaged Compass skill for Claude Code, Cursor, or Codex.
  </Card>

  <Card title="API Reference" icon="code" href="/v2/api-reference">
    Browse the complete API reference documentation.
  </Card>
</CardGroup>