How does this skill help with Langfuse integration?

It provides standardized code patterns for singleton clients, trace hierarchies, and decorators to ensure your LLM observability is robust, maintainable, and idiomatic.

Can I track user sessions with these patterns?

Absolutely. It includes patterns for linking multiple traces into a single session view using userId and sessionId parameters for longitudinal user analysis.

Does it support both TypeScript and Python?

Yes, the skill includes specific patterns and best practices for both TypeScript (manual instrumentation) and Python (using the @observe decorator system).

How should I handle Langfuse client shutdown?

The skill recommends a singleton pattern with a dedicated shutdownAsync function to ensure all batched data is flushed to the server before the application process exits.

Langfuse SDK Patterns

Name: Langfuse SDK Patterns
Author: jeremylongshore

byjeremylongshore

•

1,242

•

Analytics & Monitoring

Implements idiomatic Langfuse SDK patterns and best practices for comprehensive LLM application observability and tracing.

This skill provides a comprehensive set of implementation patterns for the Langfuse SDK, enabling developers to integrate robust observability into their LLM applications. It covers essential architectural decisions such as singleton client management, trace hierarchy construction (traces, spans, and generations), and nested operations for complex RAG workflows. By following these production-ready patterns for both TypeScript and Python, developers can ensure consistent data collection, session tracking, and user feedback integration, leading to better debugging and optimization of AI-driven systems.

Key Features

01Python decorator-based observability for clean and non-intrusive code instrumentation

021,242 GitHub stars

03Integrated session tracking, user identification, and evaluation scoring

04Singleton client patterns for consistent SDK instantiation and resource management

05Comprehensive trace lifecycle management for synchronous and asynchronous operations

06Hierarchical nesting of spans and generations for complex multi-step LLM workflows

Use Cases

01Setting up user feedback loops and performance scoring for production AI models

02Implementing full-stack observability in a TypeScript or Next.js LLM application

03Adding automated tracing to Python-based RAG pipelines using @observe decorators

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jeremylongshore/claude-code-plugins-plus-skills langfuse-sdk-patterns

For use in Claude.ai and ChatGPT

Download Skill