Usage¶
You can run the server using Task, which simplifies the process, or with Docker Compose for a containerized environment.
Running the Server Locally¶
To run the FastAPI server on your local machine:
The server will be available at http://0.0.0.0:8080. It will automatically reload when code changes are detected.
Running with Docker¶
To build the server in a Docker container:
To run the server in a Docker container:
The server will be available at http://localhost:8080.
compose.yaml
---
services:
mcp-server:
build:
context: .
args:
VERSION: ${GIT_TAG:-latest}
image: nicholaswilde/mcp-server:${GIT_TAG:-latest}
ports:
- "8080:8080"
volumes:
- ./agents-library:/app/agents-library
environment:
- PUID=${PUID:-1000}
- PGID=${PGID:-1000}
- SERVER_PORT=${SERVER_PORT:-8080}
- AGENTS_LIBRARY_PATH=${AGENTS_LIBRARY_PATH:-/app/agents-library}
- MCP_SERVER_NAME=${MCP_SERVER_NAME:-mcp-server}
- MCP_SERVER_STREAMABLE_HTTP_PATH=${MCP_SERVER_STREAMABLE_HTTP_PATH:-/}
- MCP_SERVER_JSON_RESPONSE=${MCP_SERVER_JSON_RESPONSE:-true}
restart: unless-stopped
healthcheck:
test: ["CMD", "curl", "-f", "http://localhost:8080/health"]
interval: 30s
timeout: 10s
retries: 3
start_period: 5s
Available Tasks¶
task: Available tasks for this project:
* build: Build and push the Docker image for multiple architectures (amd64, arm64).
* compile: Compile requirements.in to update requirements.txt.
* default: List all available tasks.
* docker-run: Run the Docker container.
* export: Export task list
* gflash: Run Gemini with the gemini-2.5-flash model.
* glite: Run Gemini with the gemini-2.5-flash-lite model.
* gpro: Run Gemini with the gemini-2.5-pro model.
* install: Install Python dependencies using pip-sync.
* install-pre-commit: Install pre-commit hooks for the repository.
* lint: Run linting and formatting checks using ruff.
* pre-commit-run: Run all pre-commit hooks on all files.
* ruff: Run ruff checks and formatting.
* run: Run the FastAPI server locally.
* serve: Serve mkdocs-material
* test: Run Python unit tests using pytest.