Getting Started

What is Getting Started?

Getting Started is a project that tests the initial LiteLLM MCP implementation by integrating with Go-based MCP servers and clients from the mark3labs library.

How to use Getting Started?

Install Go 1.24.1 and uv, then create a Python 3.13 virtual environment and install bridge dependencies. Run the Go MCP server on SSE transport (port 8080), and separately run the Go client connecting to that server’s URI. Optionally use Docker to start the LiteLLM proxy via the bridge’s docker-compose file.

Key features of Getting Started?

• Tests the initial LiteLLM MCP implementation
• Uses Go-based MCP server and client (mark3labs)
• Supports SSE transport over HTTP
• Includes Docker setup for LiteLLM proxy
• Provides both Python SDK and Go client examples
• Requires Go 1.24.1 and uv

Use cases of Getting Started?

• Verifying LiteLLM MCP integration with Go components
• Experimenting with MCP server/client communication over SSE
• Learning how to set up an MCP development environment
• Testing interoperability between LiteLLM and third-party MCP tools

FAQ from Getting Started

What dependencies are required?

Go 1.24.1 and uv must be installed; Python 3.13 is used for the bridge.

How do I run the Go MCP server?

Use go run main.go -t sse -p 8080 from the server directory to start the server on SSE transport.

How do I connect the Go client to the server?

From the client directory, run go run main.go -mcpUri 'http://localhost:8080/sse'.

How can I use Docker instead of the Python bridge?

Run cd bridge && docker compose up to start the LiteLLM proxy via Docker.

What transport does the example use?

The example uses SSE (Server-Sent Events) over HTTP; the server is started on port 8080.