Langfuse
Langfuse is the open-source LLM engineering platform: tracing, evaluations, prompt management and cost analytics for AI apps. Its hosted MCP server lets agents discover and fetch managed prompts from your Langfuse project directly.
Langfuse is an open-source platform for building and improving LLM applications, covering tracing and observability, evaluations, prompt management, datasets and cost analytics. The hosted MCP server exposes Langfuse Prompt Management to agents, so a coding agent or assistant can list, search and fetch your versioned, production-managed prompts on demand instead of hardcoding them. Langfuse Cloud runs in the EU by default (Frankfurt), with US and other regions available, and the MCP server authenticates with your Langfuse project API keys.
- Category
- AI Infrastructure & LLM Tools
- EU hosting
- Yes
- HQ
- Germany
- Pricing
- Freemium
- Open source
- Yes
Teams running LLM apps who manage prompts, traces and evals in Langfuse and want agents to pull live prompts.
Teams that do not use an LLM observability or prompt-management workflow.
- LLM tracing and observability
- Prompt management
- Evaluations and datasets
- Cost and usage analytics
- EU data residency
Fetch versioned production prompts into an agent; manage prompt updates without redeploys; ground assistants in your prompt library.
Works with any MCP client (Claude, Cursor, VS Code) and through the Tulimoa gateway.
Tulimoa has tested and verified this tool's MCP server. Agents can plug in with confidence.
- Endpoint
- Authentication
- API key