Munge is an external MCP server that gives your AI agents direct access to company records, geocoding, weather, macro data, and public filings across Norway, Europe, and the US. Structured, versioned, and available as ready-to-use tools. No custom integrations, no scraping, no per-source API key setup. One URL. Ten data sources.
Each source is exposed as MCP tools your agent can call directly. No custom parsing, no schema wrangling, no separate auth workflows per provider.
Munge works with any MCP-compatible client. Pick yours and paste the exact config shown below.
Take the JSON block above and paste it into your client MCP config file.
All data is open and free. No API key, no signup, no billing details.
Your agent now has 28 tools across 10 sources. Ask about weather, filings, macro indicators, geocoding, and more.
Filter all 28 tools to see exactly what your agent can access. Every tool has a full JSON schema and an LLM-friendly description.
A real agent, a real prompt, and a multi-tool response built with Munge.
You can build and maintain this stack yourself, or connect once and focus on product logic.
External MCP means no local Docker, no package setup, and no custom bridge code. Your client connects over HTTP.
Most connected sources are public APIs. Munge keeps them unified behind a single MCP endpoint.
When upstream APIs evolve, we absorb interface changes so your prompts and tool routing can stay stable.
Each tool has a clean name, typed JSON schema, and model-friendly description. We handle source-specific payload formats.
Public information should be accessible. Start free, ship fast, and self-host if you want full control.
For experimentation, demos, and weekend projects. No signup required.
For teams building production agents with cross-market data. Full catalog included.
Deploy your own instance on Railway, Fly.io, or your own infrastructure. Full control over traffic and operations.
Get running in under a minute. No signup, no credit card, no install. Paste the URL into your MCP client and start querying public data.