Crypto MCP Server provides three tools for interacting with cryptocurrency market data via the MCP protocol. The server exposes get_ticker for live price and market summary data, get_ohclv for historical OHLCV candles, and stream_ticker as an asynchronous real-time price stream. You can query real-time data for supported symbols and exchanges, fetch historical candles with configurable timeframes, and subscribe to a live stream of price updates that yields new data at a regular interval. This setup includes in-memory caching, error handling, and a lightweight MCP server architecture to facilitate manual testing and integration with MCP clients.

To use the tools locally, run the server using the provided entry point (python -m server.main) from the repository root or the server directory. Once running, you can invoke the tools via the MCP client of your choice, or use the included client_test.py to exercise the tools manually. The tests cover symbol validity, exchange support, API errors, and streaming behavior, ensuring reliable operation across common scenarios.

Prerequisites:

• Python 3.10+ (recommended 3.10 or newer)
• Git
• Internet access to install dependencies from requirements.txt

Step-by-step installation:

1. Clone the repository: git clone https://github.com/yourusername/crypto-mcp-server
2. Navigate to the project directory: cd crypto-mcp-server
3. Create and activate a virtual environment (optional but recommended): python -m venv venv

   On Windows:

   venv\Scripts\activate

   On macOS/Linux:

   source venv/bin/activate
4. Install dependencies: pip install -r requirements.txt
5. (Optional) Run tests to verify setup: pytest -vv

Notes:

• The server is Python-based and uses asyncio for streaming. Ensure network access to any required exchange APIs via ccxt.
• You can adjust environment-specific settings by editing the server/configuration as needed.

Tips and common issues:

• If you encounter import errors, make sure you are running from the project root or adjust PYTHONPATH to include the project directory.
• The in-memory cache is TTL-based; cached results expire after the configured duration to balance freshness and API load.
• For streaming, ensure the event loop is running and that your MCP client properly iterates the async generator from stream_ticker.
• If tests fail due to API rate limits or network issues, consider running tests with mocks or local fixtures where applicable.
• The project uses ccxt for exchange APIs; some exchanges may require API keys or rate-limiting considerations. Validate your target symbol and exchange are supported in tests.