Tool overview
Langfuse is listed under AI Governance Safety & Evaluation AI tools.
What is Langfuse?
Langfuse is an open-source AI engineering platform for tracing, evaluations, prompt management, datasets, experiments, and annotation. Teams can use Langfuse Cloud or self-host the MIT-licensed core with the same APIs and data model, while enterprise security features use commercial licensing.
Best for
AI engineering teams combining tracing, evaluations, prompts, and self-hosting
Who is it for?
Decision note
Best for teams that need open-source LLM observability and evaluations with a smooth choice between managed cloud and self-hosting.
Key features
LLM and agent tracing with OpenTelemetry support
Evaluations, datasets, experiments, and annotation queues
Prompt management and playground workflows
Cloud and unrestricted MIT-licensed core self-hosting
Use cases
Monitor production LLM and agent traces
Manage and version prompts
Run evaluations and experiments
Self-host observability data and workflows
Pros
- Hobby cloud plan is free
- Core paid plan starts at $29 per month
- All core self-hosted features are MIT licensed
Limitations
Observability and evaluations depend on correct instrumentation, representative data, and meaningful metrics. Self-hosted teams must operate storage, ClickHouse, upgrades, backups, and security.
Pricing details
Billing options
Pricing note
Langfuse Cloud begins with a free Hobby plan. Core costs $29 per month, Pro $199 per month, and Enterprise $2,499 per month before usage. All core self-hosted features are free under MIT; enterprise security add-ons are commercial.
Supported languages
- English
Integrations
OpenTelemetry
OpenAI and Anthropic
LangChain and LlamaIndex
Agent frameworks
Data and analytics stores
Please log in to join the discussion.