Skip to main content

Langfuse

Category: AI Infrastructure

An open-source LLM engineering platform designed for observability, metrics, and evaluations to help teams build, monitor, and optimize production-grade AI applications. Langfuse was founded in 2022. The company is led by Marc Klingen. Based in Berlin, Germany. Team size: 10-20. Total funding raised: Acquired by ClickHouse (Jan 2026). Latest round: Acquired by ClickHouse (Jan 2026) - LLM observability platform. Key investors include Lightspeed Venture Partners, General Catalyst, Y Combinator, La Famiglia, J12 Ventures.

Founded
2022
Headquarters
Berlin, Germany
Team size
10-20
Total funding
Acquired by ClickHouse (Jan 2026)

Value proposition

Provides end-to-end visibility into LLM application performance, allowing developers to debug complex traces, manage prompts, and automate evaluations to improve model reliability and reduce costs.

Products and solutions

LLM Tracing & Observability (Nested traces, latency, and cost tracking), Prompt Management (Version control and collaborative prompt engineering), Evaluation Engine (Automated LLM-as-a-judge and manual human-in-the-loop scoring), LLM Datasets & Testing (Regression testing and benchmarking frameworks), Analytics Dashboard (Usage metrics, quality trends, and token consumption monitoring)

Unique value

The leading open-source alternative to proprietary LLM monitoring tools, offering a 'developer-first' approach with deep SDK integrations and the flexibility of self-hosting for data privacy-conscious enterprises.

Target customer

AI engineers, LLM developers, DevOps teams, and enterprise software organizations building generative AI products.

Industries served

Artificial Intelligence, Software Development, Enterprise SaaS, Fintech, Healthcare & Life Sciences

Technology advantage

Leverages a high-performance architecture (built on ClickHouse) to handle massive volumes of trace data with low latency; features a modular open-source core that encourages rapid community-driven integration with the evolving AI stack.

How they differentiate

Langfuse differentiates through its 'open-source first' (MIT license) and self-hostable architecture, allowing enterprises to maintain data privacy. Unlike proprietary competitors, it offers a modular suite combining tracing, prompt management, and evaluations built on a high-performance ClickHouse-native backend.

Main competitors

LangSmith (LangChain), Arize Phoenix, Weights & Biases (W&B)

Key partnerships

ClickHouse (Parent company/Integration partner), Y Combinator (W23 Alumnus), Major AI Frameworks (Native integrations with LangChain, LlamaIndex, and LiteLLM), Model Providers (OpenAI, Anthropic, and AWS Bedrock)

Notable customers

Tines, Encord, Whale, Various YC startups

Major milestones

Accepted into Y Combinator W23 batch, Raised $4M Seed round led by Lightspeed and General Catalyst in late 2023, Reached 20,000 GitHub stars in 2024, Acquired by ClickHouse in January 2025 to integrate AI observability into the ClickHouse data stack

Growth metrics

20,000+ GitHub stars, 6M+ Docker pulls, and reached $1.1M ARR as of mid-2024.

Market positioning

The leading open-source alternative in the LLM observability and engineering (LLMOps) space, targeting developers and privacy-conscious enterprises.

Geographic focus

Global, with a strong presence in Europe (Berlin-based) and North America (via Y Combinator network).

Patents and IP

No registered patents disclosed; the company utilizes an Open Source (MIT License) strategy to drive adoption and market standard-setting.

About Marc Klingen

Marc Klingen is the Co-founder and CEO of Langfuse. He is a serial entrepreneur who previously co-founded Blair (YC S19), a fintech startup. His professional background includes serving as a Consultant at McKinsey & Company and gaining venture and tech experience at Google and Cherry Ventures. He holds a Master of Science in Management from the London School of Economics and a Bachelor's degree from WHU – Otto Beisheim School of Management.

Official website: