Langfuse

Langfuse is an open source platform for observability and evaluation of LLM applications and AI agents. It gives engineering teams a single place to trace what their application did, test changes and improve quality over time. The main capabilities are hierarchical tracing of LLM calls, tool invocations and retrieval steps, filterable by user, session, cost, latency or metadata; evaluation using LLM as a judge, heuristic functions or human review; prompt management with versioning, deployment and rollback; a playground for testing prompts on real inputs; datasets and experiments; human annotation queues; and cost and latency dashboards with alerts. It works with any language that supports OpenTelemetry, has native Python and TypeScript SDKs, and lists more than 100 integrations with frameworks and model providers. Langfuse can be used as a hosted cloud service or self hosted. Cloud pricing is published in four plans. Hobby is free and includes 50k units per month, 30 days of data access and a two seat limit. Core costs $29 per month with 100k units, 90 days of data access and unlimited users. Pro costs $199 per month with 3 years of data access and higher rate limits. Enterprise costs $2,499 per month with audit logs, SCIM and support SLAs. Additional usage is $8 per 100k units on the paid plans, and a Teams add on is $300 per month. Discounts are listed for startups, education and open source projects.