> For the complete documentation index, see [llms.txt](https://docs.algenta.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.algenta.ai/help/ai-access.md).

# Using these docs with AI

These docs are built to be read by machines as well as people. Algenta is a deep product, so the documentation is structured so that an AI assistant — Claude, ChatGPT, Cursor, Codex, Google Gemini, NotebookLM, or your own bot — can pull in accurate, current context and answer questions about the engine without guessing.

Everything here works against the published Algenta docs site (for example `https://docs.algenta.ai`). Swap in your own docs domain if you are reading a self-published copy.

## LLM-ready context files

GitBook publishes the entire documentation set as plain-text context files at well-known paths, regenerated on every change. Point any model at them.

| Path             | What it contains                                                                                                                 |
| ---------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| `/llms.txt`      | A structured map of the docs — every page title, description, and link. The right size for a system prompt or a retrieval index. |
| `/llms-full.txt` | The full text of every page concatenated into one document. Drop it straight into a long-context model.                          |
| `<any-page>.md`  | The raw Markdown of a single page — append nothing special; the page is served as Markdown to clients that ask for it.           |

```bash
curl -sS https://docs.algenta.ai/llms-full.txt -o algenta-docs.txt
# Then attach algenta-docs.txt to your prompt, or chunk it into a vector store.
```

{% hint style="info" %}
`llms.txt` is the index; `llms-full.txt` is the corpus. For a retrieval pipeline, embed `llms-full.txt` in chunks and keep `llms.txt` as the table of contents. For a single long-context call, `llms-full.txt` alone is usually enough.
{% endhint %}

## The built-in MCP server

GitBook exposes these docs as a [Model Context Protocol](/sdks/mcp.md) server, so an MCP-aware client can search and read the documentation as a live tool — no scraping, always current.

```
https://docs.algenta.ai/~gitbook/mcp
```

Add it to an MCP client the same way you would any other server. For example, in a client that takes a URL-based server:

```json
{
  "mcpServers": {
    "algenta-docs": {
      "url": "https://docs.algenta.ai/~gitbook/mcp"
    }
  }
}
```

Once connected, the assistant can answer "how do I run the Mojo engine?" or "what is `device_identity_mismatch`?" by reading the actual pages instead of relying on its training data.

{% hint style="warning" %}
The docs MCP server is distinct from Algenta's own [MCP server](/sdks/mcp.md), which exposes the *engine* (run simulations, query data) rather than the documentation. Use the docs MCP for questions about Algenta; use the product MCP to actually call the engine.
{% endhint %}

## Per-page actions

Every page carries actions for moving its content into a chat:

* **Copy as Markdown** — copy the clean Markdown of the current page to paste into any assistant.
* **Open in ChatGPT / Claude** — open the current page as context in a new chat with one click.
* **View as Markdown** — see the raw `.md` behind the rendered page.

These are the fastest way to ask a question about the page you are already reading.

## Point your own chatbot at the docs

To ground a custom assistant, you have three good options, smallest effort first:

1. **Long-context prompt** — fetch `/llms-full.txt` and attach it to the conversation. Best for one-off or low-volume use.
2. **Retrieval (RAG)** — chunk and embed `/llms-full.txt`, and use `/llms.txt` as a citation map. Best for a production assistant over a large surface.
3. **MCP tool** — connect the assistant to `/~gitbook/mcp` so it reads pages on demand. Best when you want always-fresh answers with no re-indexing.

Because the context files and the MCP server regenerate from the same source as this site, every approach stays in sync with the published docs.

## Search and AI-ingestion checks

For Google Search and Google AI surfaces, treat the docs like a public developer portal, not an app shell:

* Keep every page reachable from the GitBook sitemap and from a normal link path.
* Keep `robots.txt` crawlable for the public docs pages.
* Keep page titles, frontmatter descriptions, and first paragraphs specific enough that Google can understand the page without executing product commands.
* Keep `/llms.txt`, `/llms-full.txt`, and per-page `.md` output free of non-public filesystem paths, local usernames, raw credentials, and draft operator notes.
* Use the docs MCP endpoint only for documentation questions; do not mix it with the product MCP server that can execute engine tools.

{% hint style="info" %}
Gemini, NotebookLM, and Google-style retrieval systems work best when the answer is visible in the Markdown source and not only implied by navigation. If a page explains a setup path for Cursor, Codex, or another agent client, include the exact config file, command, environment variables, expected result, and troubleshooting path on that page.
{% endhint %}

## For documentation owners

A few of the AI features above depend on settings in the GitBook organization, not on anything in this repository:

* **Page Actions** must be enabled for the site to turn on *Copy as Markdown*, *Open in ChatGPT/Claude*, and the docs MCP server at `/~gitbook/mcp`.
* **AI Search** (in-site natural-language search) requires a GitBook Premium or Ultimate plan.
* `/llms.txt`, `/llms-full.txt`, and per-page `.md` are served automatically on any published site — no configuration needed.

## Related pages

{% content-ref url="/pages/9l598Hwbo4aGEvKi96R8" %}
[MCP server](/sdks/mcp.md)
{% endcontent-ref %}

{% content-ref url="/pages/mUN6X4yj3DJmXqbTs2Kv" %}
[Quick reference](/help/quick-reference.md)
{% endcontent-ref %}

{% content-ref url="/pages/bdxaaDSHzNrLgMm0q7zu" %}
[Glossary](/help/glossary.md)
{% endcontent-ref %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.algenta.ai/help/ai-access.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.