The AI Stack
Sign in

Langfuse

Open-source LLM engineering platform for observability, tracing, evals, and prompt management.

Updated April 2026

Overview

Segment
LLM Observability & Tracing

Product overview

Langfuse builds an open-source platform to help teams develop, monitor, evaluate, and debug LLM applications with features like traces, prompt management, evaluations, datasets, and a playground. It is used by over 40,000 builders including 19 Fortune 50 companies like Merck and Khan Academy, and was acquired by ClickHouse in January 2026. Its distinction lies in being fully self-hostable with the same codebase as the cloud version, MIT-licensed core, broad framework integrations, and enterprise-grade scalability.

Revenue model

Open-source (MIT license, self-host free); Cloud SaaS tiers: Hobby (free, 50k units/mo), Core ($29/mo, 100k units), Pro ($199/mo, 100k units + compliance), Enterprise ($2,499/mo, custom); overage $8/100k units (volume discounts); discounts for startups/OSS/edu.

Moat

Langfuse's key competitive moat is its dominant position as the most widely adopted open-source LLM observability platform, evidenced by 23,423 GitHub stars, 23.1M+ monthly SDK installs, 6M+ Docker pulls, and usage by 19 Fortune 50 and 63 Fortune 500 companies, creating strong network effects and high switching costs through deeply integrated production tracing data. This is further fortified by its January 2026 acquisition by ClickHouse, leveraging the latter's high-performance analytics database for superior scale in handling LLM data workloads while maintaining an open, extensible architecture.