Open-source LLM observability. Trace every prompt, track costs, evaluate quality. The analytics platform for AI applications.
If you're shipping LLM-powered features to production without observability, you're flying blind. Langfuse gives you request tracing, cost tracking, and quality evaluations in one place — and it's open-source, so you can self-host or use the cloud version.
Who it's for: Engineering teams building production AI applications. Especially those using LangChain or LlamaIndex who need visibility into multi-step agent workflows.
Automatically traces every LLM call — prompts, completions, latency, tokens, cost. Follows multi-step chains across agents and tools. One-line integration with LangChain, LlamaIndex, and vanilla SDKs.
Run automated evals on model outputs. Use LLM-as-judge, regex, or custom scorers. Compare models, prompts, and configurations over time with built-in dashboards.
Real-time cost breakdowns by model, user, feature, and time period. Set budgets and alerts. The CFO-friendly way to manage AI spending across your organization.
Version-control prompts outside your codebase. A/B test prompt variations. Roll back to previous versions instantly without redeploying your application.
Langfuse gives you traces, costs, and evaluations in one place. It's free to start, open-source, and integrates in minutes. Every AI application team should have this or something like it. The prompt versioning alone justifies the setup time.
LangChain's official observability platform. Deep integration with LangChain ecosystem.
LLM proxy with built-in logging, caching, and rate limiting.
AI evaluation platform. Test and improve LLM outputs systematically.
Framework for building LLM applications. Chains, agents, and tools.