# Framo Knowledge MCP Server

Framo exposes its knowledge base to AI agents over the **Model Context Protocol (MCP)**, so an assistant can read Framo's docs and answer questions about how the app works. The server speaks MCP over streamable HTTP (JSON mode) at:

```
https://framo.app/api/mcp
```

Point any MCP-capable client at that URL — no key needed for public access.

## Tools

- **`framo_list_docs`** — list the available docs (id, title, description, type). Start here.
- **`framo_search_docs`** — search the knowledge base by keyword or question; returns the most relevant docs.
- **`framo_get_doc`** — fetch the full markdown of one doc by its id (e.g. `product/agent`).

A typical flow: search or list to find a doc id, then get the doc.

## Access tiers

- **Public (anonymous)** — connect with no credentials and you get Framo's public documentation: concepts, features, and step-by-step playbooks.
- **Full (signed-in)** — a Framo account can unlock the complete bundle. While signed in, fetch a personal token from `/api/mcp/token` and send it as `Authorization: Bearer <token>` from your MCP client. The same endpoint serves the public tier when no token is present.

## Connecting

Most clients take a server URL (and optional headers). For example:

```json
{
  "mcpServers": {
    "framo": { "url": "https://framo.app/api/mcp" }
  }
}
```

The machine-readable doc index also lives at [/llms.txt](https://framo.app/llms.txt), and every public doc is browsable under [/docs](https://framo.app/docs). See [The Agent (Skills Chat)](/docs/product/agent) for Framo's in-app agent.