Langfuse
Open-source LLM observability — traces, evals, and prompt management for AI apps
Langfuse GmbH · Infrastructure
Launched 2023
Quick Answer
Langfuse is open-source llm observability — traces, evals, and prompt management for ai apps, made by Langfuse GmbH. It offers a free tier with paid upgrades. Key uses: LLM app debugging, Agent trace inspection, Prompt versioning and A/B testing.
Last updated:
Users
6K+ GitHub stars; tens of thousands of developers; widely adopted in LLMOps space
What is Langfuse?
Langfuse is an open-source LLM engineering platform for observability, prompt management, and evaluation of AI applications. It captures detailed traces of every LLM call, tool use, and retrieval step in your agent pipelines, enabling debugging, latency analysis, cost tracking, and regression testing. Integrates natively with LangChain, LlamaIndex, OpenAI SDK, and any custom LLM app via a simple decorator pattern.
What's new in Langfuse in 2026?
Langfuse v3 (2025): redesigned trace UI; annotation queues for human labeling; Prompt Experiments for A/B testing; LLM-as-judge evaluators; 6K+ GitHub stars (self-hosted); Langfuse Cloud EU and US regions; $30M+ funded; widely adopted as the open-source alternative to LangSmith
How much does Langfuse cost?
Hobby (Cloud)
$0
- 50K observations/mo
- Tracing, prompt management
- Core evaluations
Pro (Cloud)
$59/mo
- 1M observations/mo
- Annotation queues
- Experiments
- Priority support
Enterprise
Custom
- Unlimited observations
- SSO/SAML
- Self-hosted option
- SLA
- Dedicated support
Self-Hosted
$0
- MIT license
- Full features
- No data leaves your infra
What can you do with Langfuse?
- LLM app debugging
- Agent trace inspection
- Prompt versioning and A/B testing
- LLM cost monitoring
- Regression testing for AI outputs
Pros
- Open-source with self-hosting option
- Native integrations with all major frameworks
- Detailed multi-step agent traces
- Built-in evaluation datasets
- Prompt playground and versioning
- SOC 2 compliant cloud
Cons
- Cloud free tier has usage limits
- Self-hosting requires infrastructure management
- UI can be complex for simple use cases
What are the best alternatives to Langfuse?
Compare Langfuse Head-to-Head
Frequently Asked Questions about Langfuse
How much does Langfuse cost?
Langfuse pricing: Hobby (Cloud) at $0; Pro (Cloud) at $59/mo; Enterprise at Custom; Self-Hosted at $0.
Is Langfuse free?
Langfuse offers a free tier with limited usage. Paid plans unlock additional features.
What is Langfuse used for?
Langfuse is used for: LLM app debugging, Agent trace inspection, Prompt versioning and A/B testing, LLM cost monitoring, Regression testing for AI outputs.
What are the best alternatives to Langfuse?
Top alternatives to Langfuse include Weights & Biases.
What platforms does Langfuse support?
Langfuse is available on: Web (cloud), Self-hosted (Docker), Kubernetes.
Details
Platforms
Integrations
Ideal For
- LLM app developers
- AI teams debugging agents
- LLMOps engineers
- Teams needing prompt version control