Langfuse: Open Source Agent Evals & Observability

Get Started with Tracing
Step-by-step guide to ingesting your first trace using OpenAI, LangChain, or the SDKs.
Get Started with Tracing
This guide walks you through ingesting your first trace.
Langfuse joins ClickHouse
Our goal continues to be building the best AI engineering platform
Community Q&A threads
Restrict models per AI Gateway key
Call Google Vertex AI models through the AI Gateway
Evaluate with Gemini decision models
Terraform (AWS, GCP, Azure)
Langfuse launches AI Gateway
How Teams in Regulated Industries Get AI Agents into Production
Langfuse October Update
Used by 21 of the Fortune 50
110B+ observations/month
100,000+ engineers building on Langfuse
Used by 21 of the Fortune 50
110B+ observations/month
100,000+ engineers building on Langfuse
Used by 21 of the Fortune 50
110B+ observations/month
100,000+ engineers building on Langfuse
Langfuse is one open platform for every AI call your company makes and every agent your teams build. Track usage and costs, trace and evaluate agents, and use production logs to ship better AI. Cloud or self-hosted.
Instrument your agents with the SDKs or OpenTelemetry, or route model calls through the Langfuse AI Gateway without changing code. Every call lands in Langfuse for tracing, evals, and cost tracking, and you can export it to your warehouse.
Coding and desktop assistants
Support, voice, RAG, workflows
Low-code and workflow builders
Instrument your agents with the Langfuse SDKs or any OpenTelemetry exporter.
Gateway keys, model access, and cost per user for OpenAI and Anthropic calls, all traced automatically. Or bring your own gateway.
Coding and desktop assistants
Support, voice, RAG, workflows
Low-code and workflow builders
Instrument your agents with the Langfuse SDKs or any OpenTelemetry exporter.
Gateway keys, model access, and cost per user for OpenAI and Anthropic calls, all traced automatically. Or bring your own gateway.
Langfuse is the platform. The Langfuse Academy is how you, and your agents, learn what to do with it: the AI engineering loop of tracing, monitoring, datasets, experiments, and evals, with worked examples from real applications. Free, open, and readable as Markdown for coding agents.
One integrated platform to trace, manage prompts, evaluate, and experiment from prototype to production scale.
Swap the base URL and API key to route your OpenAI and Anthropic calls through Langfuse. Gateway keys show cost per user and key, and every call is traced automatically.
Hierarchical traces capture every LLM call, tool invocation, and retrieval step. Filter by user, session, cost, latency, or custom metadata.
LLM-as-a-judge, heuristic functions, or human review. Run evaluators on production data or during experiments.