Open-source LLM engineering platform for tracing, evaluation and prompt management.

Do you work at Langfuse? Claim this page for free to update its info and collect reviews.

Claim this page

About Langfuse

Langfuse is an open-source platform that helps teams build and improve applications based on large language models. It records hierarchical traces of model calls, tool use and retrieval steps, then lets teams evaluate outputs with model-based judges, custom checks or human review. A prompt management layer separates prompts from code with versioning and a playground, and datasets and experiments let teams compare changes before release. Cost and latency dashboards track usage. Langfuse can be self-hosted under an MIT license or used as a managed cloud service, and joined ClickHouse in 2026.

Listed inFoundation Models & LLM APIsasLLM Evaluation Platforms

How teams use Langfuse

No success cases yet.

Media

No media yet
Screenshots and product tours appear here once the vendor claims this page.

Features

Trace LLM calls and tool use
Evaluate outputs with LLM judges
Collect human annotations
Manage and version prompts
Run experiments on datasets
Monitor cost and latency
Route model calls through a gateway
About
Founded2023
HeadquartersBerlin, Germany
Employees10 – 50
langfuse.com