What is the Frankfurter MCP server?

The Frankfurter MCP server is a tool that exposes the Frankfurter API (a currency exchange rate API) to AI agents operating within the Model Context Protocol (MCP). It allows AI agents to access real-time and historical currency data.

What is the Frankfurter API?

Frankfurter is a free and open-source API that provides access to the latest currency exchange rates, historical data, and time series published by sources such as the European Central Bank.

What is the Model Context Protocol (MCP)?

MCP is an open protocol that standardizes how applications provide context to LLMs (Large Language Models), enabling AI agents to interact with external data sources and tools.

How do I install the Frankfurter MCP server?

You can install it using uv or Docker. Detailed instructions are in the project’s README file on GitHub.

What environment variables can I configure?

You can configure variables like HTTPX_TIMEOUT, HTTPX_VERIFY_SSL, MCP_SERVER_TRANSPORT, and FRANKFURTER_API_URL using environment variables.

What transport modes are supported?

The server supports stdio, sse (Server-Sent Events), and streamable-http transport modes.

What tools are exposed by the server?

The server exposes tools such as get_supported_currencies, get_latest_exchange_rates, convert_currency_latest, get_historical_exchange_rates, and convert_currency_specific_date.

How can I see the detailed schema for a specific tool?

You can use the tools-info command from the command-line interface (cli). The server must be running in streamable-http or sse transport mode.

Can I use this with LlamaIndex?

Yes, you can use the llamaindex-tools-list command to display the names of the tools, demonstrating LlamaIndex MCP client compatibility. You’ll need to install the LlamaIndex MCP client first (uv sync --extra opt).

How do I contribute to the project?

Install pre-commit and ruff, then enable pre-commit by running pre-commit install in the working directory. Pull requests are welcome.

How do I run the tests?

Execute uv run --group test pytest tests/. Make sure the environment variable MCP_SERVER_TRANSPORT is set to either sse or streamable-http (or not set at all).