Mixpeek MCP Server

A lightweight, production-friendly Model Context Protocol (MCP) server for Mixpeek that:

• Auto-loads Mixpeek's OpenAPI spec and exposes endpoints as MCP tools
• Supports local and hosted use (bring-your-own API key)
• Injects Authorization and X-Namespace headers
• Includes rate limits, timeouts, and redacted logs
• Ships with Docker and simple configuration

Quickstart

1. Install

Python (PyPI):

# global (recommended)
pipx install mixpeek-mcp

# or in a venv
python -m venv .venv && source .venv/bin/activate
pip install mixpeek-mcp

npm (Node):

# run once
npx @mixpeek/mcp

# or install globally
npm i -g @mixpeek/mcp

Homebrew (macOS):

# tap and install (once the tap is live)
brew tap mixpeek/tap https://github.com/mixpeek/homebrew-tap
brew install mixpeek-mcp

# temporary local formula (dev/testing):
brew install --build-from-source Formula/mixpeek-mcp.rb

2. Configure (env or your MCP client secret store)

cp env.sample .env
# edit as needed

Env vars:

• MIXPEEK_API_KEY: Your Mixpeek API key (optional if endpoints don't require auth)
• MIXPEEK_API_BASE: Default https://api.mixpeek.com
• MIXPEEK_OPENAPI_URL: Defaults to <API_BASE>/openapi.json (or /docs/openapi.json)
• MIXPEEK_NAMESPACE: Optional namespace value to send via X-Namespace
• MCP_RATE_MAX_CALLS: Default 20 per MCP_RATE_PER_SECONDS
• MCP_RATE_PER_SECONDS: Default 10
• MCP_CONNECT_TIMEOUT: Default 5
• MCP_READ_TIMEOUT: Default 30

3. Run locally (stdio)

# PyPI
mixpeek-mcp

# from source
python server.py

# npm
mixpeek-mcp

Your MCP client (e.g., Claude Desktop) can attach to this server via stdio.

Docker

docker build -t mixpeek-mcp:latest .
docker run --rm -it \
  -e MIXPEEK_API_KEY=sk_... \
  -e MIXPEEK_NAMESPACE=your_namespace \
  mixpeek-mcp:latest

Or pull and run (once published):

docker run --rm -it \
  -e MIXPEEK_API_KEY=sk_... \
  -e MIXPEEK_NAMESPACE=your_namespace \
  ghcr.io/mixpeek/mcp:latest

Distribution

• npm: @mixpeek/mcp → https://www.npmjs.com/package/@mixpeek/mcp
• PyPI: mixpeek-mcp → https://pypi.org/project/mixpeek-mcp/ (publish pending)
• Docker Hub: mixpeek/mcp (pending push), GHCR: ghcr.io/mixpeek/mcp (pending push)
• Homebrew: mixpeek/tap/mixpeek-mcp (tap repo to be created)

Submit to Docker MCP Registry

We prepared registry.json compatible with the Official Docker MCP Registry. To submit:

1. Fork the registry and create a new entry under the appropriate directory per their CONTRIBUTING guide.
2. Include our registry.json (update Docker image reference if you publish under a different org/tag).
3. Open a PR. Upon approval, it will appear in the MCP catalog and Docker Desktop's MCP Toolkit.

Configuration

Required for write-protected endpoints:

• MIXPEEK_API_KEY: Authorization: Bearer <key>

Optional:

• MIXPEEK_NAMESPACE: sets X-Namespace for isolation
• MIXPEEK_API_BASE: defaults to https://api.mixpeek.com; testing: https://server-xb24.onrender.com
• MIXPEEK_OPENAPI_URL: defaults to <API_BASE>/openapi.json (or /docs/openapi.json)
• MCP_RATE_MAX_CALLS / MCP_RATE_PER_SECONDS: simple token bucket
• MCP_CONNECT_TIMEOUT / MCP_READ_TIMEOUT: request timeouts

How it works

• Loads OpenAPI spec and maps GET/POST JSON endpoints to tools using operationId
• Tool arguments accept top-level query parameters or a query/body envelope
• Forwards requests to Mixpeek with configured headers
• Provides small allowlist (configurable) and redacts secrets in logs

Testing

Unit tests:

pip install -r requirements.txt
pytest -q

Live test (optional):

export LIVE_MIXPEEK_API_KEY=sk_...
export LIVE_MIXPEEK_OPENAPI_URL=https://server-xb24.onrender.com/docs/openapi.json
export LIVE_MIXPEEK_API_BASE=https://server-xb24.onrender.com
pytest -q tests/test_live_integration.py

Using with MCP clients

• This server uses MCP stdio transport and the official server interface, compatible with common MCP clients.
• Start the server, then point your MCP client to the stdio command (mixpeek-mcp).

References

• Mixpeek docs: https://docs.mixpeek.com/overview/introduction
• Mixpeek OpenAPI: https://api.mixpeek.com/docs/openapi.json
• MCP overview: https://modelcontextprotocol.io/docs/getting-started/intro

Notes

• For production hosting, front with HTTPS, add SSO/session issuance, per-tenant rate limits, and audit logs without bodies/headers. Keep local stdio as the default; hosted HTTP/SSE can be added later.