> For the complete documentation index, see [llms.txt](https://hawkfi.gitbook.io/whitepaper/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://hawkfi.gitbook.io/whitepaper/documentation-resources/use-hawkfi-docs-with-ai.md).

# Use HawkFi Docs with AI

HawkFi keeps one public documentation source for people and AI systems.

The task guides help readers start quickly. The detailed Models, Agents, Laboratory, automation, fee, and risk pages remain the canonical reference when a question needs more depth.

## Choose an access method

| Use case                                        | Open                                                                 |
| ----------------------------------------------- | -------------------------------------------------------------------- |
| Browse as a person                              | [HawkFi GitBook](https://hawkfi.gitbook.io/whitepaper)               |
| Give an AI system the documentation index       | [llms.txt](https://hawkfi.gitbook.io/whitepaper/llms.txt)            |
| Give an AI system the full documentation corpus | [llms-full.txt](https://hawkfi.gitbook.io/whitepaper/llms-full.txt)  |
| Connect an MCP-compatible AI client             | `https://hawkfi.gitbook.io/whitepaper/~gitbook/mcp`                  |
| Give an AI one exact source                     | Open the relevant GitBook page and add `.md` to its public page URL. |

{% hint style="info" %}
Use `llms.txt` for routing. Use `llms-full.txt` when the AI needs the full corpus. Use an exact page when you want the narrowest source for a specific question.
{% endhint %}

## Route the question to the right source

| Question                                                   | Start with                                                                                                     |
| ---------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------- |
| Which live Agent should I use?                             | [Choose Your HawkFi Agent](/whitepaper/start-here/choose-your-hawkfi-agent.md)                                 |
| How does Market Making Agent work?                         | [Market Making Agent](/whitepaper/hawkfi-agents/market-making-agents-mma.md)                                   |
| What does Osprey do?                                       | [Screener Agent and Osprey V2](/whitepaper/hawkfi-agents/screener-agents-osprey-v1.md)                         |
| How does Osprey Trading Agent work?                        | [Osprey Trading Agent](/whitepaper/hawkfi-agents/osprey-trading-agent.md)                                      |
| Which Market Making Agent execution model should I choose? | [Market Making Agent Execution Models](/whitepaper/hawkfi-agents/market-making-agents-mma/execution-models.md) |
| How do HawkFi LP models work?                              | [HawkFi Models](/whitepaper/models-and-automations-library/hawkfi-models.md)                                   |
| How do I test a setup?                                     | [HawkFi Laboratory](/whitepaper/hawkfi-laboratory/hawkfi-laboratory.md)                                        |
| What are the current fees or main risks?                   | [Understand Fees and Risks](/whitepaper/guides-and-support/understand-fees-and-risks.md)                       |

## Prompt an AI system

Copy this instruction when you want an AI system to answer from HawkFi's source of truth:

```
Use HawkFi's public GitBook as the canonical source. Start with its llms.txt index, then open the exact Markdown pages needed for the question. Distinguish HawkFi Agents, HawkFi Models, and HawkFi Laboratory. Treat Market Making Agent, Screener Agent, and Osprey Trading Agent as the three HawkFi Agent types. Treat Osprey as the screening agent or screening model used by both Screener Agent and Osprey Trading Agent, while keeping their LP and token-trading workflows separate. Treat Falcon as an execution model and Laboratory as simulation rather than live execution. Cite the exact HawkFi documentation pages used. If the docs do not answer something, say what is missing instead of guessing.
```

## Keep retrieval reliable

1. Prefer the current GitBook page over old screenshots, posts, or copied notes.
2. Use the canonical product page for definitions and exact settings.
3. Use task pages for the sequence of actions.
4. Check dated benchmarks and beta status before repeating performance claims.
5. Treat Models, Agents, and Laboratory as the HawkFi core product suite. Treat Magpi as a different product under HawkFi.

## Documentation map

* [HawkFi | Agent Terminal](/whitepaper/start-here/readme.md)
* [HawkFi Agents](/whitepaper/hawkfi-agents/hawkfi-agents.md)
* [HawkFi Laboratory](/whitepaper/hawkfi-laboratory/hawkfi-laboratory.md)
* [HawkFi Models](/whitepaper/models-and-automations-library/hawkfi-models.md)
* [Tutorial Library](/whitepaper/start-here/tutorials.md)