IRLFirst physical meetup — Bengaluru, Sat May 23, 4PM · RSVP on Luma
HomeToolsMCPHow It WorksStoriesPhilosophyCommunityArchitectureStar on GitHub
All Tools
L
MonitoringFreemiumOpen Source

LANGFUSE

Open-source LLM observability and engineering platform

MIT

ABOUT

LLM applications and AI agents are complex, non-deterministic, and difficult to debug in production. Teams lack visibility into agent behavior, cannot track costs and latency across models, and have no systematic way to evaluate output quality or safely iterate on prompts.

INSTALL
pip install langfuse

INTEGRATION GUIDE

1. Trace and debug LLM agent workflows with hierarchical spans and OpenTelemetry support 2. Track token usage, cost, and latency across models and providers in production 3. Evaluate LLM outputs using automated LLM-as-a-judge, heuristic, or human annotation workflows

TAGS

observabilityllmtracingevaluationprompt-managementopentelemetry
Langfuse — AI Tool | Agentic AI For Good