tools
Langfuse
Langfuse traces, evaluates, and monitors AI agents and LLM applications, with prompt management, experiments, and cost analytics.
In inglese
Langfuse is an open-source platform for tracing LLM calls, tool use, retrieval steps, and agent workflows. Teams use it to inspect production behavior, manage prompts, run evaluations, compare experiments, and track quality, cost, and latency.
It is available as Langfuse Cloud or self-hosted software. Cloud pricing includes a free Hobby plan, while usage beyond included units and some enterprise capabilities cost extra. Some features depend on external LLM APIs.
Features
- Capture hierarchical traces for LLM calls, tools, retrieval, sessions, and agents
- Manage, version, cache, compose, and deploy prompts separately from application code
- Run LLM-as-a-Judge, code, human, and custom evaluations
- Create datasets and compare prompt, model, and code changes with experiments
- Monitor cost, latency, volume, quality, and scores in dashboards
- Use public APIs, SDKs, CLI, and MCP Server to query and manage data
- Self-host with Docker, Kubernetes, AWS, Azure, or GCP
- Open-source software under the MIT license
Use cases
- Debug multi-step AI agent traces in production
- Evaluate chatbot quality with live traces and test datasets
- Compare prompt or model changes before deployment
- Track LLM costs and latency by user, session, feature, or model
- Review outputs manually and build labeled evaluation datasets
- Export trace and evaluation data to external systems
Pros
Cons
Latest updates
- Langfuse v4 is live: faster at scale, with more ways to search, monitor, and evaluate
Initial table loads drop from seconds to milliseconds, and dashboards over longer time ranges load at least 10x faster.
Capabilities
- Command line — “Langfuse CLI Full API access from the terminal for agent workflows, scripts, and CI/CD.” source
- Choice of models — “Test prompts on real production inputs and compare models side-by-side.” source
- Self-hosted — “Self-host at scale” source
- API — “REST APIs for everything” source
- Official SDKs — “Best-in-class SDKs: Native SDKs for Python and JavaScript” source
- Traces and evaluates — “Trace, evaluate, and improve AI agents with one open platform.” source
Get it
Security
- SOC 2 Type II — “The production service is based on the same open-source Langfuse codebase, is covered by SOC 2 Type II and ISO 27001 audits, and undergoes annual third-party penetration tests.” source
- SOC 2 — “We take active steps to demonstrate our commitment to data security and privacy such as annual SOC2 Type 2 and ISO27001 audits as well as External Penetration Tests.” source
- ISO 27001 — “The production service is based on the same open-source Langfuse codebase, is covered by SOC 2 Type II and ISO 27001 audits, and undergoes annual third-party penetration tests.” source
- GDPR — “Langfuse is GDPR compliant, and offers data retention, data masking and data deletion capabilities to manage the processing of personal data .” source
- HIPAA — “We offer a DPA , provide a HIPAA-ready region , and adhere to our Privacy Policy .” source
- Data kept in the EU — “Langfuse Cloud – fully-managed SaaS (multi-tenant) with US, EU, Japan, and HIPAA data regions” source
Pricing
- Starting price
- $29/mo
- Prices checked
- 2026-09-25
Hobby
Free
- 50k units / month included
- 30 days data access
- 2 users
- Community support via GitHub
- All platform features (with limits)
Core
- $29 / month
- $8/100k units
- 100k units / month included
- 90 days data access
- Unlimited users
- In-app support
- Everything in Hobby
Pro
- $199 / month
- $8/100k units
- 100k units / month included
- 3 years data access
- Data retention management
- Unlimited annotation queues
- High rate limits
- Prioritized in-app support
Enterprise
- $2499 / month
- $8/100k units
- 100k units / month included
- Audit Logs
- SCIM API
- Custom rate limits
- Uptime SLA
- Named lead support engineer