Perfox Docs MCP (for AI agents)

Perfox publishes its documentation as a remote MCP server so AI assistants — Claude, Cursor, Copilot, ChatGPT, or your own agent — can search and read these docs at inference time. This is the fastest way to give a coding agent accurate, up-to-date knowledge of the Perfox platform.

This is read-only docs search. It is not the Integrations described in MCP in Perfox — that's how your agent calls your business logic. This page is about querying our docs.

Endpoint

https://perfox.ai/mcp

Transport: Streamable HTTP (JSON-RPC 2.0 over POST), stateless — no auth, no session.
Read-only, CORS-open, free to use.

Tools

Tool	Arguments	Returns
`search_docs`	`query` (string), `limit` (int, default 5, max 20)	Ranked pages — title, URL, snippet. Start here.
`get_doc`	`path` (e.g. `concepts/byop`) or a full URL	The full Markdown of one page.
`list_docs`	—	The whole table of contents as Markdown links.

Connect from Claude / Cursor

Add it as a remote MCP server:

json

{
  "mcpServers": {
    "perfox-docs": {
      "url": "https://perfox.ai/mcp"
    }
  }
}

Then ask, e.g. "Using the perfox-docs tools, how do I embed the widget with identity verification?" — the agent calls search_docs, then get_doc.

Prefer plain files? Use `llms.txt`

If your tool reads static LLM context instead of MCP, point it at:

https://perfox.ai/llms.txt — index / table of contents
https://perfox.ai/llms-full.txt — the entire docs in one file
Any page also has a clean Markdown twin — append .md (e.g. https://perfox.ai/docs/concepts/byop.md).

Quick test

bash

curl -sX POST https://perfox.ai/mcp \
  -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call",
       "params":{"name":"search_docs","arguments":{"query":"BYOP credentials"}}}'

Perfox Docs MCP (for AI agents) ​

Endpoint ​

Tools ​

Connect from Claude / Cursor ​

Prefer plain files? Use llms.txt ​

Quick test ​