hugoles/langfuse-mcp

What it does

hugoles/langfuse-mcp is a Model Context Protocol (MCP) server that integrates with the Langfuse platform, providing comprehensive access to observability and monitoring data for large language model (LLM) applications. It exposes 27 read-only tools covering traces, observations, sessions, scores, score configurations, prompts (with version and label support), datasets, dataset items, dataset runs, metrics, models, projects, comments, media, and health checks.

How to set up

The tool is distributed as an npm package and can be installed via npx. Users need to configure their Langfuse API credentials and then instantiate the MCP server. The package is written in TypeScript and can be integrated into any MCP-compatible client application such as Claude, ChatGPT, or AI Studio.

Key features

• 27 read tools covering all major Langfuse data types
• Support for traces, observations, and session monitoring
• Prompt management with version and label tracking
• Dataset and dataset run management
• Score configuration and metrics access
• Multi-platform compatibility (ChatGPT, Claude, AI Studio, Anthropic, MCP)
• Distributed as installable npm package via npx