> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.caesar.xyz/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.caesar.xyz/_mcp/server.

# llms.txt

Caesar's documentation supports [llms.txt](https://llmstxt.org/), a standard for exposing documentation to AI developer tools like Cursor, GitHub Copilot, ChatGPT, and Claude.

## What is llms.txt?

When AI coding assistants access documentation, they typically receive HTML pages full of navigation, styling, and other markup. This wastes tokens and reduces context quality.

`llms.txt` solves this by providing:

* **Markdown instead of HTML**: 90%+ reduction in token usage
* **Structured summaries**: Each page distilled to a one-sentence description
* **Hierarchical organization**: AI tools can understand your documentation structure

## Available endpoints

| Endpoint         | Description                                         |
| ---------------- | --------------------------------------------------- |
| `/llms.txt`      | Lightweight summary with page descriptions and URLs |
| `/llms-full.txt` | Complete documentation content in Markdown          |

> **Tip**
>
> Use `/llms.txt` for discovery and navigation. Use `/llms-full.txt` when you need the AI to understand specific implementation details.

## Using with AI tools

### Cursor

Add Caesar's documentation to your Cursor context:

1. Open Cursor settings
2. Navigate to **Features** → **Docs**
3. Add `https://docs.caesar.xyz/llms.txt`

Now you can reference Caesar documentation in your prompts with `@Caesar`.

### Claude / ChatGPT

Paste the llms.txt URL directly in your conversation:

```
Read https://docs.caesar.xyz/llms-full.txt and help me integrate Caesar's research API into my Python application.
```

### GitHub Copilot

Reference the documentation in your comments:

```python
# See https://docs.caesar.xyz/llms.txt for Caesar API documentation
# Implement a function that creates a research job with auto mode
```

## Query parameters

Filter the output to reduce tokens further:

| Parameter     | Description                        | Example             |
| ------------- | ---------------------------------- | ------------------- |
| `lang`        | Include only specific SDK language | `?lang=python`      |
| `excludeSpec` | Exclude OpenAPI specification      | `?excludeSpec=true` |

**`Python-only documentation`**

```bash title="Python-only documentation"
curl https://docs.caesar.xyz/llms.txt?lang=python
```

**`Exclude API spec`**

```bash title="Exclude API spec"
curl https://docs.caesar.xyz/llms-full.txt?excludeSpec=true
```

## Hierarchical access

Access llms.txt at different documentation levels:

```
/llms.txt                              # Full documentation summary
/documentation/get-started/llms.txt    # Get Started section only
/api-reference/llms.txt                # API Reference only
```

> **Info**
>
> The comprehensive `/llms-full.txt` is only available at the root level.

## Example output

**`llms.txt format`**

```text title="llms.txt format"
# Caesar API Documentation

> Caesar is an agentic research API that transforms questions into comprehensive, citation-backed answers.

## Get Started

- [Introduction](/): Overview of Caesar's research capabilities
- [Authentication](/documentation/get-started/authentication): API key generation and usage
- [Rate Limits](/documentation/get-started/rate-limits): Concurrency and monthly limits

## Integration

- [SDKs](/integration/sdks): Python and TypeScript client libraries
- [OpenAI Compatibility](/integration/openai-compatibility): Use OpenAI SDK with Caesar
...
```

## Best practices

#### Start with the summary

Use `/llms.txt` first to help the AI understand what's available, then drill into specific pages.

#### Filter by language

If you're working in Python, use `?lang=python` to exclude irrelevant code examples.