Langfuse
Open-source tracing, prompt management and evaluation for LLM products.
ABOUT
What is Langfuse?
Langfuse is an open-source LLM engineering platform that covers the full development lifecycle: tracing LLM calls, managing prompt versions, running evals and analyzing usage metrics. It integrates with every major LLM framework including LangChain, LlamaIndex, OpenAI SDK and Vercel AI SDK through either native integrations or a thin wrapper. The self-hosted edition runs on Docker and is production-grade.
BEFORE YOU SIGN UP
What you should know
COMMUNITY
From the forums
Langfuse has rapidly become the most recommended open-source LLM observability tool on GitHub. The combination of a well-documented self-hosted option and a functional free cloud tier is frequently cited as the reason teams choose it over proprietary alternatives. The 30-day retention limit on the cloud free tier and the complexity of setting up evals pipelines are the most common criticisms.
— Community sentiment
SEE ALSO · AI & machine learning
Alternatives to Langfuse
ABOUT US
Honest, independent, no fluff.
No paid placements. Just a clear look at what this does, what it costs, and what to know before you commit.
FAQ
Questions, answered.
Does Langfuse work with non-OpenAI models?＋
Yes. Langfuse is model-agnostic. It traces any LLM call through its SDK wrapper. You can also use LiteLLM as a proxy to normalize calls from dozens of model providers.
How hard is it to self-host Langfuse?＋
The Docker Compose setup works out of the box for most teams. A Kubernetes helm chart is also available for production deployments. The main requirements are a Postgres database and enough memory for the worker processes.