How does this skill authenticate with Langfuse?

It uses HTTP Basic Auth with your Langfuse Public and Secret keys, which are typically stored in the project's .env.local file.

Langfuse is an open-source LLM observability platform used for tracing, evaluation, and analytics of AI applications.

Can I use this with a self-hosted Langfuse instance?

Yes, the skill supports custom endpoints for self-hosted deployments in addition to Cloud US, EU, and HIPAA regions.

What kind of data can I ingest with this skill?

You can create traces, spans, generations, and events to map out your entire LLM request lifecycle for better debugging.

How do I query specific metrics?

The skill uses the GET /api/public/v2/metrics endpoint to retrieve aggregated data based on views like observations or numeric scores.

Langfuse Observability

Name: Langfuse Observability
Author: drdanmaggs

bydrdanmaggs

0•

Analytics & Monitoring

Integrates Langfuse LLM observability to track traces, manage evaluation scores, and monitor AI application performance via curl.

The Langfuse skill enables Claude to interact directly with the Langfuse API, providing a seamless way to implement and manage LLM observability within your development environment. It facilitates the ingestion of traces and spans, the retrieval of detailed observations, and the management of evaluation scores to ensure high-quality AI outputs. Whether you are debugging production traces, tracking model latency, or analyzing cost metrics, this skill streamlines the process of integrating robust monitoring into your AI-native workflows.

Key Features

01Add and update numeric, categorical, or boolean evaluation scores

02Support for Cloud EU, US, HIPAA, and self-hosted regional endpoints

030 GitHub stars

04Access aggregated analytics and metrics for cost and model performance

05Ingest traces, spans, and generations via the Langfuse Ingestion API

06Query and retrieve detailed observations with advanced filtering and pagination

Use Cases

01Debugging LLM logic by retrieving specific traces and spans from production

02Implementing automated evaluation systems by posting evaluation scores from CLI tests

03Monitoring AI infrastructure costs and token usage through aggregated metrics queries

Key Features

01Add and update numeric, categorical, or boolean evaluation scores

02Support for Cloud EU, US, HIPAA, and self-hosted regional endpoints

030 GitHub stars

04Access aggregated analytics and metrics for cost and model performance

05Ingest traces, spans, and generations via the Langfuse Ingestion API

06Query and retrieve detailed observations with advanced filtering and pagination

Use Cases

01Debugging LLM logic by retrieving specific traces and spans from production

02Implementing automated evaluation systems by posting evaluation scores from CLI tests

03Monitoring AI infrastructure costs and token usage through aggregated metrics queries