Arize Phoenix
AI observability and evaluation platform
Discover. Evaluate. Test.
Open source evaluation and observability platform for AI agents
Langfuse is an open platform to trace, evaluate and improve AI agents and LLM applications. It records traces of each request, including model calls, retrieval steps and tool use, so that production behaviour can be understood and issues fixed.
Evaluations can be run with LLM-as-a-judge, user feedback, manual annotation or custom scores, and experiments compare prompt or model versions on datasets. Prompt management, a playground and metrics on quality, cost and latency complete the platform.
Langfuse is available as a cloud service with a free tier, or can be self-hosted. Most of the code is open source under the MIT licence, with some enterprise features under a commercial licence. The project is now part of ClickHouse.
There are no reviews yet.