Open sourceTypeScript · ≈38k stars · Other
Langfuse
Open-source LLM engineering platform: traces, evals, prompts.
Langfuse is the open-source observability layer for LLM apps — tracing, prompt management, datasets, evals and cost tracking — deployable self-hosted or as a cloud.
Pros & cons
Each point cites its sources and says what kind of evidence backs it. A second reviewer tried to refute every claim before it was kept.
Pros
- User reportsOpen source, self-hostable and OpenTelemetry-based, so users say it slots into existing tooling and the observability part is easy to switch to or from.[1][2]
- User reportsUsers credit it with a broader feature set (evals, prompt management) than Logfire, in one place alongside tracing.[1][2]
- User reportsUsers praise turning production generations into datasets to test edge cases.[1][2]
- User reportsCommonly recommended over LangSmith once teams mix frameworks or need self-hosting for compliance.[1][2]
Cons
- User reportsSelf-hosting v3 needs ClickHouse (a painful v2-to-v3 migration and disk-hungry).[1][2]
- MetadataNot pure MIT: code under the ee/ directories uses a separate commercial licence, and ClickHouse Inc. now holds the copyright after the Jan 2026 acquisition.[1][2]
- User reportsTrace ingestion can lag by minutes when using older SDKs/endpoints, per users and a maintainer reply.[1][2]
- User reportsIntegration bugs: Gemini token/cost mismatches, OTel conflicts with Google ADK, and GCS media upload failures when self-hosted.[1][2][3]
- User reportsThe MCP server only supports static API keys, not SSO/OIDC (a heavily requested issue).[1]
- Best for
- Teams wanting open-source, self-hostable, OTel-native LLM tracing plus evals across frameworks.
- Watch out for
- ClickHouse operations when self-hosting, ee-licensed features, and SDK-dependent ingestion delay.
Assessed Oct 29, 2026 · adversarially reviewedSuggest a correction
Activity
- Last push
- · yesterday
- Latest release
- v4.83.0
· 2d ago - Last 90 days
- 20+ releases
- · 2d ago
What's Changed Features feat(topics): new Scatter view, remove List by @nmtrang29 in https://github.com/langfuse/langfuse/pull/20355 feat(annotation-queues): navigate between items from single queue item view (LFE-18664) by @nkabardin in…
- · 4d ago
What's Changed Features feat(cloud): gate legacy v3 surfaces behind cutoff switches by @maxdeichmann in https://github.com/langfuse/langfuse/pull/20054 feat(api-keys): add Developer preset and rename Ingestion to Application by @marksalpeter in…
- · 5d ago
What's Changed Features feat(datasets): filter dataset items by status by @marliessophie in https://github.com/langfuse/langfuse/pull/20186 feat(ai-gateway): add Google Vertex AI as a gateway provider by @wochinge in https://github.com/langfuse/langfuse/pull/20079 feat(trace):…
- · 8d ago
What's Changed Features feat(clickhouse): send query parameters in the request body by default by @Steffen911 in https://github.com/langfuse/langfuse/pull/20049 feat(ai-gateway): add Google AI Studio as a gateway provider by @wochinge in…
- · 9d ago
What's Changed Features feat(ai-gateway): add OpenRouter as a pre-release gateway provider by @wochinge in https://github.com/langfuse/langfuse/pull/20002 feat: add SUBSTITUTE_OK no-data mode for monitors by @marksalpeter in https://github.com/langfuse/langfuse/pull/19981…
Checked against GitHub on Nov 21, 2026. Release notes are shortened to plain text.
The team behind it
Berlin, Germany · Est. 2023
Open-source LLM engineering platform: tracing, prompts, evals.
Langfuse is an open-core platform for observing and improving LLM apps — trace-level debugging, prompt management, datasets and evals. Self-hostable, with a cloud offering. Acquired by ClickHouse in January 2026 as part of its push into AI observability; the open-source core remains actively maintained.
Acquired by ClickHouse (Jan 2026)
Neighbours on the workbench
Go · ≈55k stars · Apache-2.0
alibaba/open-code-review
Alibaba's hybrid deterministic-plus-agent code review CLI.
Maintained by Alibaba
v1.13.13releasedPython · ≈39k stars · MIT
stanfordnlp/dspy
Programming, not prompting: optimize the whole pipeline.
Maintained by Stanford NLP
3.5.0releasedTypeScript · ≈30k stars · Elastic-2.0
mastra-ai/mastra
The TypeScript framework for production agents.
Maintained by Mastra
@mastra/core@1.84.0releasedConnections
Relations from the Harness Hub knowledge graph, as charted in the State of Harness reports.
Maintained by · 1
- Langfuse (ClickHouse)Startup
Similar entries · 3
- Arize AIStartup
- Langfuse (ClickHouse)Startup
- PhoenixProject
