mcp -vanmoof

The VanMoof MCP Server implements the Model Context Protocol (MCP) to bridge AI agents with VanMoof services. It exposes a set of tools that let agents fetch customer data, city ride data, rider preferences, and ride statistics from VanMoof accounts. Agents can query for bike details, locations, weekly ride summaries, and city/world ride statistics, enabling richer, context-aware interactions. The server declares a collection of tools such as get_customer_data, get_vanmoof_cities, get_rider_preferences, get_rides_summary, get_rides_for_week, get_city_rides_thisweek, and get_world_rides_thisweek. To use these capabilities, start the MCP server and connect via the MCP Inspector or your preferred MCP client, then issue prompts that map to these tools (e.g., asking for bike details, or requesting weekly ride statistics). The inspector helps test and debug by simulating MCP requests and showcasing structured responses from the VanMoof API via the MCP interface.

Prerequisites:

• Python must be installed on your system.
• Access to the VanMoof API credentials (username and password).
• Optionally install the UV toolchain for faster development (recommended per README): uv (Python project manager) from https://docs.astral.sh/uv/.

Install steps:

1. Clone the repository: git clone https://github.com/stefanstranger/mcp-server-vanmoof
2. Change into the project directory: cd mcp-server-vanmoof
3. Ensure you have a Python environment (recommended: virtualenv or conda) and install dependencies if required (refer to requirements.txt if present).
4. Create and configure environment variables for VanMoof credentials:
   • VANMOOF_USERNAME: your VanMoof account email
   • VANMOOF_PASSWORD: your VanMoof account password Example (bash): source myenv/bin/activate export VANMOOF_USERNAME="your_email@example.com" export VANMOOF_PASSWORD="your_secure_password"
5. Start the MCP server using Python (as shown in the README examples): python server.py
6. Optional: launch MCP Inspector to test interactions:
   • Start MCP Inspector via MCP CLI or npx as guided in the README to verify tool commands and responses.

Notes:

• If you prefer the uv/uvx workflow described in the README, you can adapt the startup to use uv run --with mcp[cli] ... mcp run server.py with the appropriate environment variables.

Tips and common issues:

• Ensure VANMOOF_USERNAME and VANMOOF_PASSWORD are kept secure; avoid sharing your environment with credentials exposed.
• If the MCP Inspector cannot authenticate, double-check credentials and ensure network access to VanMoof API.
• When testing weekly or city statistics, confirm the date formats accepted by get_rides_for_week and related methods.
• If running locally via Python, ensure you are in the project root so server.py can resolve any relative paths.
• If you migrate to a containerized or packaged setup, replicate the environment variables and expose necessary ports as required by your MCP client.

Environment variables and configuration options:

• VANMOOF_USERNAME: VanMoof account email
• VANMOOF_PASSWORD: VanMoof account password
• Any other API-related secrets should be provided through environment variables or a secrets manager per your deployment environment.