Langfuse
Open-source LLM observability, tracing and evals
Overview
Langfuse is an open-source LLM engineering platform for traces, evaluations, prompt management, datasets, cost tracking and dashboards. The core is MIT-licensed and self-hostable, with enterprise modules under a commercial key; Langfuse Cloud runs in US, EU and Japan regions from a free Hobby plan to $2,499 a month. ClickHouse acquired Langfuse in January 2026.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- All platform features, with limits
- 30 days data access
- 2 users
- Community support
- 90 days data access
- Unlimited users
- In-app support
- 3 years data access
- Data retention management
- SOC 2 and ISO 27001 reports
- HIPAA-ready region
- Prioritised support
- Everything in Pro plus Teams
- Audit logs and SCIM API
- Uptime and support SLAs
- Named support engineer
- All core features and APIs
- Helm chart and deployment docs
- Community support
- Project-level RBAC
- Data retention policies and audit logs
- Support SLA
Use cases
What you can produce with Langfuse
- Traces of every LLM call, tool call and retrieval with latency, tokens and cost
- Self-hosted MIT core so telemetry stays in your environment
- Prompt management and versioning without a code redeploy
- LLM-as-judge evaluators over production traces
- Datasets and experiments comparing models or prompt versions
- Cost dashboards by user, model or feature
- User feedback scores attached to specific traces
Langfuse covers the core of LLM observability: nested traces with cost and latency, prompt versioning, datasets, experiments and LLM-as-judge evaluators, through framework-agnostic SDKs and OpenTelemetry. The MIT core self-hosts without feature limits; SCIM, audit logs and retention policies need an enterprise key.
The caveats are operational: self-hosting means running PostgreSQL, Redis, ClickHouse and object storage. Since January 2026 it belongs to ClickHouse, which has kept the licence unchanged. Cloud regions are US, EU and Japan, so self-hosting is the route to keeping data in Southeast Asia.
What we found
Langfuse is an open-source platform for observing and evaluating LLM applications. It records traces of each model call, tool call and retrieval step, tracks token cost and latency, manages and versions prompts, and runs datasets, experiments and LLM-as-judge evaluators against production data. It was started by Marc Klingen, Max Deichmann and Clemens Rawert, who went through Y Combinator's Winter 2023 batch, and it is operated by Langfuse GmbH in Berlin with a smaller San Francisco office.
The biggest change is ownership. On 16 January 2026 ClickHouse announced it had acquired Langfuse, alongside a $400 million Series D; Reuters reported a $15 billion valuation for ClickHouse. Langfuse's announcement says the product stays open source and self-hostable with no planned changes to licensing, and the repository bears that out: everything outside its enterprise directories is MIT-licensed, while SCIM, audit logs and data-retention policies need a commercial key when self-hosted. ClickHouse said at the time that Langfuse had more than 26 million SDK installs a month and was used by 19 of the Fortune 50, both company figures.
Development has stayed fast. Langfuse v4, which the company bills as real-time and much faster, is now current, and Langfuse Cloud becomes v4-only on 16 November 2026. A Tokyo cloud region opened in April 2026, joining the US and EU; there is still no Southeast Asian region.
Pricing is published. Self-hosting the open-source core is free, and the cloud Hobby plan is free with 50,000 units a month. Core costs $29 a month, Pro $199 with three years of data access and compliance reports, and Enterprise $2,499 with audit logs, SCIM and SLAs. Earlier versions of this page showed no paid tiers at all.
The costs are operational. A self-hosted deployment needs PostgreSQL, Redis, ClickHouse and object storage, which is real work for a small team, and the interface is dense. Langfuse suits engineering teams that run LLM features in production and want to own their telemetry.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
About this listing
This entry was compiled from publicly available data including Langfuse's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Langfuse unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Langfuse directly →
Spotted something out of date? Suggest an update →
Related reading
More in Code & Dev Tools
