MCP Server
Bagisto provides an optional MCP (Model Context Protocol) server that lets an AI agent search this documentation on demand, from inside the editor, while it writes code.
It is optional: /llms.txt and /llms-full.txt cover the same knowledge statically. The MCP server is worth adding when you want the agent to pull the exact current page for an endpoint mid-task instead of ingesting the whole index.
Why use it
- Live lookup while coding. The agent fetches the precise request/response shape for an endpoint at the moment it needs it — no copy-pasting docs into the chat.
- Always current. It indexes the published docs, so the agent reads the same reference you do, not a stale snapshot from its training data.
- Smaller context. Instead of loading
llms-full.txt, the agent pulls only the page it is working on, leaving room for your actual code.
For building a storefront, that means when the agent reaches the checkout step it can look up the exact address-input fields and the shipping-then-payment sequence on demand, rather than guessing or asking you.
Tools it exposes
| Tool | What it does |
|---|---|
search_api_docs | Full-text search across the API documentation |
list_endpoints | Enumerate documented endpoints (filterable by surface / menu) |
get_doc | Fetch one page's full content by path |
Install
- Repository:
bagisto/mcp
The repo README has the setup and the exact command to register the server with your agent (for example claude mcp add for Claude Code, or the editor's MCP settings for Cursor / Windsurf).
Do you need it?
No — it is a convenience, not a requirement. If you have installed the Agent Skills and the agent can reach /llms.txt, you already have everything needed to build. Add the MCP server if you want on-demand, in-editor doc search on top of that.