Queries and analyzes Langfuse trace data directly from Claude Code to debug LLM performance and token usage.
Langfuse Traces is a specialized skill for Claude Code that integrates Langfuse observability directly into your terminal workflow. It allows developers to investigate AI workflow executions, retrieve specific trace details, and monitor token consumption without needing to switch to a web dashboard. By providing instant access to spans, generations, and session history, it streamlines the debugging process for LLM-powered applications and helps optimize AI behavior through real-time data analysis of production or staging logs.
Key Features
010 GitHub stars
02Retrieve detailed trace data including specific observations and spans
03List and filter recent LLM traces with configurable limits
04Monitor session history and group traces by session IDs
05Analyze token usage and model performance metrics
06Generate compact summary views for quick snapshots of recent activity
Use Cases
01Debugging failed LLM calls or unexpected model outputs in real-time
02Investigating latency issues by inspecting granular trace spans and observations
03Analyzing token consumption and costs across different AI workflows
