The Flutter MCP Server exposes core Dart and Flutter tooling as programmable endpoints via the Model Context Protocol (MCP). This enables AI assistants, automation scripts, and CI workflows to invoke tasks such as code analysis, formatting, fixes, project scaffolding, running tests, and other Dart/Flutter commands through a consistent API layer. In addition to tool endpoints, it offers resource endpoints for documentation, news, and curated community examples, all while handling secrets and configuration through environment variables for secure operation. Use cases include automated code quality checks in CI, AI-assisted project generation, and remote tooling orchestration for Flutter projects. You can access tool endpoints to run analyze, format, fix, create, run, and test commands, and you can query resources to fetch up-to-date guidance or sample projects.

Prerequisites:

• Docker installed and running (preferred for local containerized deployment).
• Access to a terminal with Docker permissions.
• Optionally, a local Dart/Flutter SDK if you want to run without Docker (dart pub get, dart run, etc.).

Installation steps (Docker-based):

1. Build the Docker image for the Flutter MCP Server (from repository root): docker build -t flutter_mcp_server .

2. Create or prepare an environment file with required secrets and configuration keys. Example:

   .env

   Add your environment-specific variables here (e.g., API keys, secrets, endpoints)

   API_KEY=your_api_key_here SECRET_KEY=your_secret_key_here

3. Run the MCP server in a container, wiring in the environment file: docker run -it --env-file .env flutter_mcp_server

Alternative (local Dart/Flutter execution, if you prefer not to use Docker):

1. Ensure Dart SDK and Flutter SDK are installed and configured.
2. From repository root, install dependencies and start the server (example): dart pub get dart run bin/flutter_mcp_server.dart

Note: The README indicates that Smithery Cloud deployment and registry listing are not yet available, so the Docker route is the recommended method for local usage.

Environment variables: All sensitive configuration should be managed via environment variables (see .env.example in the repo). Ensure you supply keys securely and do not commit secrets to source control. If you plan to integrate with Windsurf or Cursor clients, use the provided example mcp_config.json or mcp.json snippets to start the server with Docker (or adjust the image name as needed). Docker is the recommended deployment method for reproducible and secure local or cloud-like usage. If you encounter port or network issues, verify that the Docker daemon is running and that any required ports are exposed by the hosting environment. For troubleshooting, check the server logs to confirm that tool endpoints (analyze, format, fix, create, run, test) are being invoked correctly and that environment variables are being loaded.