How does it handle complex pipelines like RAG?

It uses manual spans to wrap specific steps like embedding and vector search, linking them all under a single parent trace for clear visualization of the entire pipeline.

What is the primary purpose of this Langfuse skill?

It provides a standardized workflow for implementing end-to-end tracing and observability in LLM applications using the Langfuse platform.

Can I use this with models other than OpenAI?

Yes, it includes implementation patterns for Anthropic Claude, LangChain integration, and general-purpose manual tracing for any AI provider.

Does it support streaming responses?

Absolutely. It features specific code patterns to track and record full outputs and token usage from streaming completions.

Langfuse Core LLM Tracing

Name: Langfuse Core LLM Tracing
Author: jeremylongshore

byjeremylongshore

•

1,449

•

Analytics & Monitoring

Implements end-to-end observability for LLM applications by tracing calls, spans, and generations using Langfuse.

Langfuse Core Workflow A provides a comprehensive set of patterns for integrating Langfuse tracing into AI applications. It enables developers to monitor LLM performance, track token usage, and visualize complex multi-step chains or RAG pipelines. Whether using automatic wrappers for OpenAI or manual spans for custom logic, this skill ensures complete visibility into the execution flow, making it essential for debugging, cost management, and performance optimization of production-grade AI features.

Key Features

01Integrated token usage and cost monitoring

02Automatic OpenAI and Anthropic SDK tracing wrappers

031,449 GitHub stars

04Hierarchical trace linking for parent-child operation visibility

05Manual span management for complex RAG and agentic workflows

06Real-time tracking for streaming LLM responses

Use Cases

01Debugging complex RAG pipelines with granular span tracking

02Monitoring token consumption and latencies across distributed AI agents

03Adding observability to existing OpenAI or Anthropic production applications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jeremylongshore/claude-code-plugins-plus-skills langfuse-core-workflow-a

For use in Claude.ai and ChatGPT

Download Skill

Key Features

01Integrated token usage and cost monitoring

02Automatic OpenAI and Anthropic SDK tracing wrappers

031,449 GitHub stars

04Hierarchical trace linking for parent-child operation visibility

05Manual span management for complex RAG and agentic workflows

06Real-time tracking for streaming LLM responses

Use Cases

01Debugging complex RAG pipelines with granular span tracking

02Monitoring token consumption and latencies across distributed AI agents

03Adding observability to existing OpenAI or Anthropic production applications

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jeremylongshore/claude-code-plugins-plus-skills langfuse-core-workflow-a

For use in Claude.ai and ChatGPT

Download Skill