LinkedIn Content Creation MCP Server

# Docker Guide This guide explains how to use Docker with the LinkedIn MCP Server. ## CLI Commands The LinkedIn MCP Server provides a comprehensive CLI for running in different modes: ```bash # Run in STDIO mode (for Claude Desktop) linkedin-mcp stdio # Run in HTTP mode (API server) linkedin-mcp http --host 0.0.0.0 --port 8000 # Auto-detect best transport mode linkedin-mcp auto # Run with debug logging linkedin-mcp stdio --debug # Get help linkedin-mcp --help ``` ### Available Modes - **stdio**: Standard I/O mode for Claude Desktop integration - **http**: HTTP API server mode (requires `uvicorn` and `starlette`) - **auto**: Automatically detect and use the best transport mode ## Quick Start ### Build the Docker Image ```bash make docker-build ``` or ```bash docker build -t chuk-mcp-linkedin:latest . ``` ### Run the Container **STDIO Mode (default):** ```bash make docker-run-stdio ``` **HTTP Mode:** ```bash make docker-run-http ``` **Development Mode:** ```bash make docker-run-dev ``` ## Docker Compose Profiles The `docker-compose.yml` file defines three profiles: ### 1. STDIO Profile (Default MCP Mode) ```bash docker-compose --profile stdio up -d ``` This runs the server in standard I/O mode for integration with Claude Desktop or other MCP clients. ### 2. HTTP Profile (API Mode) ```bash docker-compose --profile http up -d ``` This runs the server in HTTP mode, exposing an API on port 8000. Access the server at: http://localhost:8000 ### 3. Development Profile ```bash docker-compose --profile dev up -d ``` This mounts the source code directory for live development. ## Environment Variables Create a `.env` file in the project root: ```env # Required - LinkedIn OAuth credentials LINKEDIN_CLIENT_ID=your_client_id_here LINKEDIN_CLIENT_SECRET=your_client_secret_here # Required for Development - Session storage SESSION_PROVIDER=memory # Optional - Server configuration MCP_SERVER_MODE=stdio DEBUG=0 ``` ## Volume Mounts The Docker containers mount two directories: 1. **`.env` file**: Read-only mount for credentials 2. **`.linkedin_drafts/`**: Persistent storage for drafts ## Docker Commands ### Build Commands ```bash # Build production image make docker-build # Build development image make docker-build-dev ``` ### Run Commands ```bash # Run in STDIO mode make docker-run-stdio # Run in HTTP mode make docker-run-http # Run in development mode make docker-run-dev ``` ### Management Commands ```bash # View logs make docker-logs # Stop containers make docker-stop # Open shell in container make docker-shell # Clean up everything make docker-clean # Build and test make docker-test ``` ## Multi-Stage Build The Dockerfile uses a multi-stage build for optimal image size: 1. **Builder Stage**: Installs all dependencies and builds the package 2. **Runtime Stage**: Contains only what's needed to run the server This reduces the final image size significantly. ## Security Features - Runs as non-root user (`mcpuser`) - Minimal runtime dependencies - No build tools in final image - Health check configured ## Health Check The container includes a health check that verifies the package can be imported: ```dockerfile HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \ CMD python -c "import chuk_mcp_linkedin; print('OK')" || exit 1 ``` ## Troubleshooting ### Container won't start ```bash # Check logs docker-compose logs # Or make docker-logs ``` ### Permission issues Ensure the `.linkedin_drafts/` directory is writable: ```bash chmod 755 .linkedin_drafts ``` ### Port already in use (HTTP mode) Change the port in `docker-compose.yml`: ```yaml ports: - "8001:8000" # Use 8001 instead of 8000 ``` ## Advanced Usage ### Custom Dockerfile Build ```bash docker build \ --build-arg PYTHON_VERSION=3.12 \ -t chuk-mcp-linkedin:custom \ . ``` ### Run with Custom Command ```bash docker run -it --rm \ -v $(pwd)/.env:/app/.env:ro \ -v $(pwd)/.linkedin_drafts:/app/.linkedin_drafts \ chuk-mcp-linkedin:latest \ python -m chuk_mcp_linkedin.async_server --port 8000 ``` ### Network Configuration All containers use the `mcp-network` network for inter-container communication. ```bash docker network inspect mcp-network ``` ## Production Deployment ### Using Docker Compose ```bash # Production mode with resource limits docker-compose -f docker-compose.yml -f docker-compose.prod.yml up -d ``` ### Using Kubernetes See `k8s/` directory for Kubernetes manifests (coming soon). ### Using Docker Swarm ```bash docker stack deploy -c docker-compose.yml linkedin-mcp ``` ## Maintenance ### Update Image ```bash # Pull latest changes git pull # Rebuild image make docker-build # Restart containers docker-compose down docker-compose --profile stdio up -d ``` ### Backup Drafts ```bash # The drafts are stored in .linkedin_drafts/ tar -czf linkedin-drafts-backup.tar.gz .linkedin_drafts/ ``` ## Resources - [Dockerfile Reference](https://docs.docker.com/engine/reference/builder/) - [Docker Compose Reference](https://docs.docker.com/compose/compose-file/) - [Multi-stage builds](https://docs.docker.com/build/building/multi-stage/)