modal

This MCP server provides a bridge for managing Modal volumes and deploying Modal applications from within Cursor. It exposes a set of tools under the Modal Volume Operations group (list, inspect, copy, upload, download, and remove) as well as a deployment tool to publish Modal apps. The server communicates with the Modal CLI and relies on UV for dependency management in the host project. To use it, point Cursor at the configured mcp.json entry for modal-mcp-server and invoke the available tool methods through Cursor's MCP interface. The output for each operation is returned in a standardized JSON format that includes success status, data payloads for read operations, and structured messaging for file operations and deployments.

Prerequisites:\n- Python 3.11+ (as required by the server)\n- UV package manager installed in your environment (uv)\n- Modal CLI configured with valid credentials in the target environment\n\nInstallation steps:\n1) Clone the repository and navigate into it:\n git clone https://github.com/smehmood/modal-mcp-server.git\n cd modal-mcp-server\n\n2) Install dependencies using uv (as per the project requirements):\n uv sync\n\n3) Ensure Modal credentials are configured in the environment where the server will run (e.g., within the virtual environment used by your project).\n4) Ensure the server.py script is accessible and update the Cursor mcp.json configuration to point to the correct absolute path for your deployment.\n\nOptional: test the server locally by running the uv command from the mcp_config example after replacing placeholder paths with real paths.

Tips and notes:\n- The mcp.json configuration must reflect the absolute path to your modal-mcp-server repository and the server.py entry point.\n- The server relies on the UV toolchain; ensure your environment has uv and the Modal CLI installed in the same Python/virtual environment.\n- For deployment-related operations, ensure the target project within UV uses UV for dependency management and that the Modal CLI is installed in its virtual environment.\n- When performing file operations, the API returns structured fields such as stdout and stderr to aid debugging.\n- If you encounter permission or path issues, verify that the paths used in --project and the server.py path are accessible by the user running UV.