# Langfuse

Open-source LLM observability, tracing & prompt management platform

Langfuse provides full-lifecycle tracing, evaluation, and monitoring for AI agents executing complex multi-step tool calls.

## Overview

- **Website**: https://langfuse.com
- **Docs**: https://langfuse.com/docs
- **GitHub**: https://github.com/langfuse/langfuse
- **License / Pricing**: Open Source
- **Directory Entry**: https://agentaccess.dk/enablers/langfuse

## About

When AI agents execute tool calling across Danish APIs, scrapers, and MCP servers, Langfuse captures complete execution traces, token consumption, latency, and tool input/output payloads. Developers can debug agent failures, track costs, and evaluate output quality in real time.

## Key Features

- Detailed distributed tracing of agent tool calls & model steps
- Self-hostable via Docker or managed cloud platform
- Cost tracking, latency monitoring, and token analytics
- Integrations with Vercel AI SDK, LangChain, LlamaIndex, and OpenAI

## Danish System Application Recipes

### Trace and Debug Danish Agent Tool-Calling Workflows
Wrap your agent in Langfuse to inspect every API call made to CVR, DAWA, or DMI, monitoring payload correctness and API latency.

## Quickstart

```typescript
import { Langfuse } from 'langfuse'

const langfuse = new Langfuse()
const trace = langfuse.trace({ name: 'danish-cvr-lookup' })
const span = trace.span({ name: 'fetch-cvr-api' })

// Perform API request...
span.end({ output: { company: 'Novo Nordisk A/S', cvr: '24256790' } })
await langfuse.shutdownAsync()
```

---
Source: [AgentAccess Enablers](https://agentaccess.dk/enablers/langfuse)