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

# AI Integration

> How AI agents and LLMs can use Argyros docs and APIs.

# AI Integration

Argyros provides resources optimized for AI agents and LLM-powered tools.

## Resources

| Resource          | URL                           | Purpose                                                                                                                                                  |
| ----------------- | ----------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **llms.txt**      | `/llms.txt`                   | LLM-optimized index of all docs pages. Use this to discover which page has the information you need.                                                     |
| **llms-full.txt** | `/llms-full.txt`              | Complete API reference in a single file. All parameters, response schemas, error tables, and code examples. Fetch this to get everything in one request. |
| **skill.md**      | `/skill.md`                   | Decision-tree playbook for AI agents. Maps user intents to API endpoints with parameter guidance, response schemas, and error recovery.                  |
| **OpenAPI spec**  | `/api-reference/openapi.json` | Machine-readable OpenAPI 3.1 spec for all endpoints. Use for code generation or structured API exploration.                                              |

## How to use

### For LLM agents (Cursor, Claude Code, Windsurf, etc.)

1. Fetch `/llms-full.txt` to get the complete API reference in one file (parameters, schemas, errors, code examples).
2. Or fetch `/llms.txt` for a page index, then read specific pages as needed.
3. Use `/skill.md` as a playbook for making API calls. It has decision trees, response schemas, and error recovery.

### For automated integrations

1. Fetch `/api-reference/openapi.json` for the machine-readable spec.
2. Read `/skill.md` for the intent → endpoint decision tree.
3. Use the parameter tables and error recovery flows to build robust integrations.

## llms.txt

The `llms.txt` file follows the [llms.txt standard](https://llmstxt.org/). It contains a one-line summary of every docs page with its URL, optimized for LLM context windows.

## skill.md

The `skill.md` file is a structured playbook that helps AI agents:

* **Route intents.** "swap SOL for USDC" → `GET /quote` then `POST /swap`
* **Choose parameters.** Which fields are required, what defaults to use
* **Handle errors.** What to do when a quote returns "no route found"
* **Compose flows.** Multi-step workflows like quote → build → sign → submit