Langfuse is an open source LLM engineering platform. It helps teams collaboratively develop, monitor, evaluate, and debug AI applications. Langfuse can be self-hosted in minutes and is battle-tested. Proudly made with ClickHouse open source database.
🪢 Open source agent evals & observability: Trace, evaluate, and improve LLM applications with one open platform.
Trace Pi coding agent sessions with Langfuse — agent turns, model generations, token usage with cache and reasoning splits, cost, tool calls, and nested subagents.
Interact with the full Langfuse API from the command line. Built for AI agents and power users.
A Langfuse integration for Openclaw to trace your conversations cleanly
🪢 Langfuse JS/TS SDKs - Instrument your LLM app and get detailed tracing/observability. Works with any LLM or framework
🪢 Langfuse Python SDK - Instrument your LLM app with decorators or low-level SDK and get detailed tracing/observability. Works with any LLM or framework
Trace Claude Code AI coding assistant conversations, user inputs, and tool usage with Langfuse for comprehensive observability.
