---
title: "Langfuse: LLM evaluation platform for builders, free tier and alternatives"
canonical_url: https://cookiesforai.app/tools/langfuse
markdown: https://cookiesforai.app/tools/langfuse.md
publisher: "CookiesForAI"
last_updated: 2026-11-11T12:00:00+01:00
type: service
summary: "Langfuse is an open source LLM engineering platform by Langfuse that provides tracing, prompt management, datasets, experiments, and online and offline evaluation."
related:
  - "https://cookiesforai.app/tools/langsmith"
  - "https://cookiesforai.app/tools/braintrust"
---

> Full index of this site: https://cookiesforai.app/llms.txt  

# Langfuse

Langfuse is an open source LLM engineering platform by Langfuse that provides tracing, prompt management, datasets, experiments, and online and offline evaluation.

## Facts

| Fact | Value |
|---|---|
| Made by | [Langfuse](https://langfuse.com/) |
| Category | [LLM evals and tracing](https://cookiesforai.app/tools/llm-evals) |
| License | Open source (MIT) |
| Source | https://github.com/langfuse/langfuse |
| Free tier | Free plan |
| Website | https://langfuse.com/ |
| Page updated | 11 NOV 2026 |
| Facts checked | 11 NOV 2026 ([how](https://cookiesforai.app/method)) |

## What Langfuse does

Langfuse is an LLM engineering platform for recording traces, tracking tokens and costs, managing prompts, and evaluating LLM applications and agents. It supports datasets and experiments for offline evaluation, plus online evaluators, user feedback, and annotation workflows.

Langfuse is available as cloud software and as an open source self-hosted product. The cloud Hobby plan includes platform features with limits, while paid plans increase data access and operational capacity for production projects.

## Who Langfuse is for

- Teams that need both LLM traces and repeatable datasets or experiments.
- Builders who want to start with a free hosted plan or self-host the open source product.
- Coding-agent workflows that benefit from a hosted documentation MCP server and agent skill.

## Who should pick something else

- Applications needing unlimited historical data on the free cloud plan, because Hobby data access is limited.
- Teams that want only basic infrastructure metrics, because Langfuse is focused on LLM application and agent work.
- Projects that cannot instrument application calls, because tracing and evaluation depend on recorded observations or supplied data.

## Langfuse: one strength, one limit

Strength: Langfuse's Hobby plan includes tracing, prompt management, datasets, experiments through the SDK, and online and offline evaluation with plan limits.

Limit: The Hobby cloud plan has 30 days of data access and supports two users.

## What Langfuse costs

Free tier: Free plan. The Hobby cloud plan is free with limits on usage, history, and users. Paid cloud plans extend retention and capacity, while Langfuse also documents free self-hosting.

Current prices: https://langfuse.com/pricing. Prices change often, so check the maker's own page before you buy.

## Langfuse with AI coding agents

Langfuse hosts an unauthenticated Docs MCP server and an authenticated project-scoped Cloud MCP server. It also publishes a Langfuse Agent Skill and setup instructions for Claude Code, Codex, Cursor, Windsurf, and Copilot in VS Code.

## Questions people ask

### Can Codex use Langfuse documentation directly?

Yes. Langfuse documents a hosted Docs MCP endpoint and provides a Codex CLI configuration command for it.

## Alternatives to Langfuse

- [LangSmith](https://cookiesforai.app/tools/langsmith): LangSmith is an agent evaluation and observability platform by LangChain that records traces, runs offline and online evaluations and compares application versions.
- [Braintrust](https://cookiesforai.app/tools/braintrust): Braintrust is an AI evaluation and observability platform by Braintrust that connects datasets, experiments, scorers, production traces and human review.

## Compare LLM evals and tracing tools

All three have a free plan.

Langfuse uses the MIT license; [Braintrust](https://cookiesforai.app/tools/braintrust) and [LangSmith](https://cookiesforai.app/tools/langsmith) are proprietary.

[Compare all 3 →](https://cookiesforai.app/tools/llm-evals)

## More in LLM evals and tracing

- [Braintrust](https://cookiesforai.app/tools/braintrust): Braintrust is an AI evaluation and observability platform by Braintrust that connects datasets, experiments, scorers, production traces and human review.
- [LangSmith](https://cookiesforai.app/tools/langsmith): LangSmith is an agent evaluation and observability platform by LangChain that records traces, runs offline and online evaluations and compares application versions.
