Langfuse
FreemiumOpen source AI engineering platform for tracing, evaluating, prompt management, and metrics to debug and improve LLM applications and AI agents.
Informations sur l’outil
Qu'est-ce que Langfuse ?
Langfuse is an open source AI engineering platform that gives teams one place to trace LLM and agent behavior, manage prompts, run evaluations and experiments, and monitor cost and latency. It connects production signals into a continuous loop so developers can understand how applications behave, test improvements, and ship higher-quality AI applications and agents with confidence. The product is MIT-licensed and available as a managed cloud service or fully self-hosted.
Fonctionnalités principales
Observability / Tracing
Hierarchical traces capture every LLM call, tool invocation, and retrieval step, with filtering by user, session, cost, latency, or custom metadata.
Evaluation
Support for LLM-as-a-judge, heuristic functions, or human review, runnable on production data or during experiments.
Prompt Management
Separate prompts from code with one-click deployments and rollbacks so teams can collaboratively improve production prompts.
Playground
Test prompts on real production inputs and compare models side-by-side.
Experiments
Define test cases, run experiments, and compare results side by side.
Human Annotation
Collaborative human-in-the-loop workflows to review traces and create golden datasets.
Cost & Latency Monitoring
Monitor cost, latency, and quality with dashboards and automated alerts.
Langfuse CLI
Command-line interface that wraps the full Langfuse API for AI coding agents and power users to create datasets, pull traces, update prompts, and automate workflows.
OpenTelemetry Native & Integrations
Works with any language and framework supporting OTel instrumentation, plus 100+ integrations including major agent frameworks and model providers.
Self-Hosting & Open Source
MIT-licensed product with self-hosting options via Docker Compose, Kubernetes (Helm), and Terraform for AWS, GCP, and Azure.
Cas d'utilisation
- #1
Debug and improve LLM applications
Use traces, evals, prompt management, and metrics to understand behavior and improve LLM application quality, cost, and latency.
- #2
Full LLM engineering loop
Connect tracing, monitoring, datasets, experiments, and evaluation so production signals drive continuous improvement of agents.
- #3
Agent and coding-agent workflows
Let coding agents (e.g. Cursor, Claude Code) manage Langfuse via CLI to create datasets, pull traces, update prompts, and run common workflows.
- #4
Production tracing for generative features
Trace and debug generative design or other AI features in production, as used by enterprise AI teams.
Plans tarifaires
Homepage offers Start free for cloud. Product is open source under MIT license with free self-hosting options (Docker Compose, Kubernetes Helm, Terraform). Exact paid cloud tiers are not clearly detailed in the provided official sources.
Alternatives
Foire aux questions
What is Langfuse?
Is Langfuse open source?
Can I self-host Langfuse?
What can I do with the Langfuse CLI?
Which frameworks and model providers does Langfuse integrate with?
How does Langfuse help improve production LLM apps?
Langfuse Contacts et réseaux sociaux
- SociétéLangfuse GmbH
- Assistance
- Site Internet :
- Twitter
- YouTube
- LinkedIn
- GitHub
Avis
Partagez votre expérience avec cet outil...