kapyn
RadarEval & observability

Langfuse

Eval & observability · Tool

Visit Website
Kapyn's take, worth it

Open-source tracing and evals, joined ClickHouse and remains open source. Self-host and own the data; more setup than Helicone, more control.

About Langfuse

Open-source tracing and evaluation for LLM applications. Joined ClickHouse and remains open source.

Open-source tracing and evals for LLM apps.

Who it's for

Engineers who want open, self-hostable observability for their AI stack.

How it works

It captures traces of model and agent runs and lets you score outputs, run evals, and manage prompts, self-hostable or managed.

Where it's used

Tracing agents, running evals, and versioning prompts in production.

CategoryEval & observability
TypeTool
Websitelangfuse.com
See Langfuse alternatives

Similar tools

LangSmithTrace, test, and debug LLM apps in production.HeliconeMonitor LLM cost, latency, and usage in one dashboard.BraintrustEvaluate and improve AI products with systematic evals.Arize PhoenixOpen-source tracing and evaluation for LLM apps.Weights & BiasesTrack experiments, models, and LLM apps in one place.ChatGPTReason, write, and prototype with OpenAI's frontier models.

Find more on the Radar

Browse the curated tools, models and MCP servers in one place.

Open the Radar