Langfuse MCP server

This is a small helper that connects your AI assistant to Langfuse, the tool teams use to watch how their AI apps behave. Once it is set up, you can ask your assistant questions about your Langfuse data in normal words instead of clicking around the dashboard. It is handy for anyone who checks traces, scores, or prompts and wants quick answers without leaving the chat.

What this MCP server does

You ask your AI something about your Langfuse data, like why a trace failed or what a prompt says. The AI passes that request to this helper program. The helper talks to Langfuse using your project keys and pulls back the matching traces, scores, prompts, or numbers. Then the AI reads that data and explains it to you in the chat. Everything it does is read only, so it looks things up but never changes them.

Click to zoom

What you can do with it

• List recent traces, filtered by level, user, or time
• Walk through a single trace and all of its observations
• Show the scores a user or session received
• Fetch a prompt by name, version, or label
• List datasets and their evaluation runs
• Summarize daily cost, latency, or token usage
• Check that your Langfuse keys work with a health ping

Try asking your AI

• “List the 5 most recent traces with level ERROR.”
• “Show me trace abc123 with all its observations.”
• “What scores did user alice@example.com receive this week?”
• “Get the production version of prompt customer-support.”

What it gives back to you

You get back plain answers in the chat: lists of traces or sessions, a breakdown of a single trace, score values, prompt text, or daily numbers for cost and usage. The AI usually summarizes what it found and can show the raw details if you ask. Nothing is written back to Langfuse, so your data stays as it is.

Before you start

Who it's for

People who work with AI apps and check Langfuse for traces, scores, prompts, or costs, such as support engineers, product folks, and developers who prefer asking questions over clicking dashboards.