Langfuse

What Langfuse does

Langfuse is an open-source AI engineering platform that provides observability, prompt management, evaluation, metrics, and human annotation for production LLM and AI agent applications. The platform captures hierarchical traces of every LLM call, tool invocation, and retrieval step via OpenTelemetry-native instrumentation and stores them on a ClickHouse OLAP backend with PostgreSQL for transactional state, Redis/Valkey for async ingestion, and S3-compatible blob storage for large multi-modal payloads, enabling millisecond analytics across millions of traces at a reported 10B+ observations per month. Its product surface includes an LLM Observability module, version-controlled Prompt Management with one-click deployments and rollbacks, an Evaluation module supporting LLM-as-a-judge and heuristic scoring, Experiments, a Playground, Human Annotation queues, Metrics dashboards with alerts, a native Model Context Protocol (MCP) server, a Public REST API, and a Query SDK.

Langfuse serves enterprise engineering teams building production AI agents and LLM applications (19 of Fortune 50, 2,300+ customers, 100,000+ engineers) as well as data-sensitive organizations self-hosting behind their firewall, and mid-market teams adopting through its open-source distribution. The company monetizes through a freemium SaaS cloud product (Free tier of 50k observations/month, usage-based Pro, quote-based Enterprise), a self-hosted Enterprise Edition sold under license, and, following its January 2026 acquisition by ClickHouse, co-distribution through ClickHouse's Cloud regions and enterprise sales channels. The MIT-licensed codebase, 29.6k GitHub stars, 300+ contributors, 100+ native integrations across model providers and agent frameworks, and compliance posture (SOC 2 Type II, ISO 27001, GDPR, HIPAA eligible) form the core of its product-led, community-led, and enterprise sales motions.

Name

Langfuse

Legal name

Langfuse GmbH

Website

https://langfuse.com

Company type

Private

Founded year

2022

Operating status

Acquired

Headcount range

11–50 employees

Short description

Langfuse is an open-source AI engineering platform that provides observability, prompt management, evaluation, and metrics for production LLM and AI agent applications, used by 19 of Fortune 50 and 2,300+ customers.

Ownership category

akta.pro rank

Product category

LLM Observability Software

NAICS

Software Publishers (5132), Software Publishers (513210)

SIC

Services-Prepackaged Software (7372)

akta.pro primary industry

LLMOps & Generative AI Platforms (Prompt/Agent Orchestration, RAG) (HDAEANAD)

akta.pro secondary industry

Prompt Engineering, Orchestration & LLMOps Tooling (HDAAACAD)

Keywords

• LLM observability platform
• AI agent tracing
• Prompt management software
• LLM evaluation tools
• AI engineering platform

Headquarters

HQ city

Berlin

HQ country

Germany

HQ region

Europe

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Marketing or Sales, Operations

Revenue model

1. Langfuse Cloud (SaaS): Hosted cloud product sold as a tiered SaaS subscription. Free tier of 50k observations/month with no credit card required, plus paid Pro and Enterprise tiers. Customers can choose between EU, US, Japan, and HIPAA-eligible US data regions.
2. Usage-based cloud billing: Cloud usage is metered by observations (billable units), allowing land-and-expand as customer trace volume grows. The free tier caps at 50k observations/month, with Pro and higher tiers scaling by consumption.
3. Self-Hosted Enterprise Edition (license): On-premises/sovereign deployment requires an Enterprise Edition license key; advanced administrative features such as Data Retention are gated to the Enterprise Edition for self-hosted customers.
4. Freemium open-source distribution: MIT-licensed open-source product with free self-hosting option acts as a funnel into paid Cloud and Enterprise Edition; the entire codebase, including all product features, is MIT-licensed and self-hostable for free.

Pricing tiers

Go-to-market motion4 records

Distribution channels4 records

Marketing channels7 records

Core offering

Langfuse is an open source AI engineering platform that provides tracing, evaluation, prompt management, and metrics for LLM applications and AI agents. It captures hierarchical traces of every LLM call, tool invocation, and retrieval step, and offers self-hostable deployment under the MIT license alongside a managed Cloud SaaS and a Self-Hosted Enterprise Edition.

Differentiator

Problem solved

Functional benefit

Products and services

• LLM Observability Hierarchical tracing module for capturing every LLM call, tool invocation, and retrieval step across AI agent and LLM applications; supports sessions, environments, sampling, masking, and multi-modality for engineering and MLOps teams.
• Prompt Management Version-control, deploy, and rollback prompts independently from application code with variables, message placeholders, prompt composability, folders, labels, caching, A/B testing, GitHub integration, and webhooks.
• Evaluation Run LLM-as-a-judge, heuristic code-based, and human review evaluators against production data or experiments with NUMERIC, CATEGORICAL, BOOLEAN, and TEXT score types.
• Metrics Dashboards and automated alerts tracking cost, latency, token usage, and quality with a Metrics API and Custom Dashboards over the ClickHouse OLAP store.
• Playground In-product testing surface that runs prompts from Prompt Management against real production inputs and compares multiple models side-by-side.
• Experiments Define dataset test cases and run experiments, comparing results side-by-side to evaluate application changes; integrates with Prompt Management and Evaluation modules.
• Human Annotation Collaborative human-in-the-loop workflows for reviewing traces and creating golden datasets via annotation queues.
• Langfuse MCP Server Native Model Context Protocol server at /api/public/mcp with read and write tools, allowing AI assistants and agents to interact with Langfuse data programmatically across EU, US, Japan, HIPAA, and self-hosted deployments.
• Langfuse Cloud Hosted SaaS offering of the Langfuse platform with regional data residency in EU, US, Japan, and HIPAA-eligible US regions; tiered subscription with a free tier (50k observations/month), Pro, and Enterprise plans.
• Self-Hosted Enterprise Edition On-premises/sovereign deployment of Langfuse requiring an Enterprise Edition license key; advanced administrative features such as Data Retention are gated to this edition for self-hosted customers.

Quantifiable outcome

• 50M+ SDK installs per month, indicating broad developer adoption.
• +3 more outcomes

Named customers12 records

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration54 records

AI capability14 records

Feature10 records

Partnerships

Twelve partnerships are on record, tiered strategic ecosystem partner (via clickhouse), flagship — acquirer and primary infrastructure partner, strategic cloud deployment partner, major hyperscaler integration, major model provider integration, major agent framework integration, no-code integration partner, major cloud + framework integration, strategic framework integration, evaluation integration partner and major cloud deployment partner.

• Confluentstrategic ecosystem partner (via clickhouse)Technology or Integration · 1 May 2026ClickHouse is deepening its ecosystem partnership with Confluent to support enterprise deployments; Langfuse benefits as part of the unified ClickHouse data platform for AI workloads.
• ClickHouseflagship — acquirer and primary infrastructure partnerStrategic or Co-development Partner · 1 January 2026ClickHouse acquired Langfuse GmbH in January 2026 concurrent with its $400M Series D funding round. The entire Langfuse team joined ClickHouse while maintaining Langfuse's open-source model, self-hosting options, and existing product contracts. Langfuse's LLM monitoring tool integrates with ClickHouse's database to track model latency and infrastructure usage; the companies plan to deepen the Langfuse–ClickHouse integration.
• Oracle Cloud Infrastructure (OCI)strategic cloud deployment partnerTechnology or IntegrationLangfuse provides observability, tracing, and evaluation for enterprise AI applications and multi-agent systems deployed on OCI using either VM or Kubernetes (OKE) patterns. Supports native OCI Object Storage integration via LANGFUSE_USE_OCI_NATIVE_OBJECT_STORAGE.
• Amazon Bedrock (AgentCore)major hyperscaler integrationTechnology or IntegrationAmazon Bedrock AgentCore has integrated Langfuse observability for detailed monitoring and debugging of AI agent performance. The implementation involves deploying Strands agents on Amazon Bedrock and configuring Langfuse for telemetry data to optimize costs and performance.
• Anthropicmajor model provider integrationTechnology or IntegrationNative model provider integration in Langfuse, including Anthropic SDK, Claude Agent SDK (Python and JS), and Claude Code tracing/dev-tool integration. Anthropic's Model Context Protocol is also natively supported.
• OpenAImajor model provider integrationTechnology or IntegrationNative OpenAI integration including the Langfuse-wrapped OpenAI client, OpenAI Agents SDK integration, and the OpenAI o3/o4-mini/o3-pro cost tracking updates.
• LangChainmajor agent framework integrationTechnology or IntegrationNative LangChain integration with a CallbackHandler for Python and JavaScript, plus LangChain DeepAgents integration and interoperability between Langfuse and LangChain traces.
• Difyno-code integration partnerTechnology or IntegrationDify ships built-in Langfuse observability and analytics; users can connect Dify applications to Langfuse for tracing and evaluation.
• Google (Vertex AI, Gemini, ADK)major cloud + framework integrationTechnology or IntegrationNative integrations with Google Gemini, Google Vertex AI, and Google ADK (Agent Development Kit) for tracing, evaluation, and prompt management.
• Vercel (AI SDK)strategic framework integrationTechnology or IntegrationNative integration with the Vercel AI SDK for tracing, evaluation, and prompt management in Next.js / Vercel deployments.
• UpTrain.aievaluation integration partnerTechnology or IntegrationUpTrain.ai integration adds 20+ evaluators to Langfuse for LLM output evaluation.
• Amazon Web Services (AWS)major cloud deployment partnerTechnology or IntegrationAWS Marketplace and first-party deployment guide for self-hosting Langfuse on AWS using Terraform, with native integration of AWS S3, KMS encryption, and ElastiCache Redis.

Scale indicators13 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• Arize AI (Phoenix): Closest direct competitor in LLM observability and evaluation. Arize Phoenix offers tracing, evaluation, and experiment tooling built on OpenTelemetry, targeting the same enterprise engineering teams that Langfuse serves, with overlapping integrations (OpenAI, Anthropic, LangChain, LlamaIndex).
• Braintrust: Direct competitor focused on LLM evaluation, experimentation, and observability. Braintrust's loop of prompts, evals, and production traces is highly comparable to Langfuse's core product, and Braintrust is routinely named alongside Langfuse in industry buyer guides.
• LangSmith (LangChain): LangChain's first-party observability and evaluation product for LLM apps. Competes head-to-head with Langfuse on tracing and evals, with the strategic advantage of being bundled into the largest agent framework (LangChain/LangGraph).
• Helicone: Open-source LLM observability proxy that provides tracing, cost monitoring, and caching for LLM calls. Targets the same developer-led adoption motion and is frequently compared directly with Langfuse in production LLM monitoring benchmarks.
• Maxim AI: Specializes in LLM evaluation, simulation, and observability for production AI agents. Overlaps with Langfuse on eval workflows and agent tracing, and is regularly listed alongside Langfuse in 2025 LLM observability buyer guides.

Broad incumbents

• Honeycomb.io: Established observability platform for high-cardinality, production tracing. While not LLM-native, Honeycomb's OTel-based tracing and event analytics are increasingly used for LLM application observability, making it a broad incumbent Langfuse is positioned against at the enterprise level.
• Datadog: Public, large-scale APM and observability incumbent with a dedicated LLM Observability / Watchdog for AI Agents product. Represents the largest competitive threat from a broad incumbent, leveraging a deep enterprise install base and a unified APM/Logs/Security platform.
• New Relic: Long-standing APM and observability platform that has added AI/LLM monitoring capabilities. Competes with Langfuse indirectly at large enterprises that already standardize on New Relic for full-stack observability.
• Dynatrace: Enterprise observability and AIOps incumbent with a unified data platform. While primarily an infrastructure and APM vendor, Dynatrace's analytics platform is increasingly used for AI workload observability and is a competitor in enterprise procurement bake-offs.

Emerging players

• WhyLabs: AI observability and data-monitoring platform with LLM evaluation, drift detection, and guardrails. Has partial overlap with Langfuse on quality monitoring and serves a similar ML/AI engineering buyer, with a comparable emphasis on data-centric AI reliability.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/langfuse
• GitHubgithub.com/langfuse/langfuse
• YouTubeyoutube.com/@langfuse
• LinkedInlinkedin.com/company/langfuse

Compliance5 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles3 records

Funding overview

Funding rounds1 record

Investors5 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Langfuse