skills.rest
avivsinai
avivsinaiCOMMUNITY
Explore all skills by this author →
𝕏@avivsinaiGitHub

langfuse

Centralize traces, exceptions, sessions, and prompts for AI debugging.

103|24|Updated Mar 21, 2025
Search Tags:#MCP#debug#observability#prompts#Langfuse#traces
One-click install
npx skills add https://github.com/avivsinai/langfuse-mcp --skill langfuse
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: langfuse
Source: https://github.com/avivsinai/langfuse-mcp/tree/main/.claude/skills/langfuse
Command: npx skills add https://github.com/avivsinai/langfuse-mcp --skill langfuse

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Langfuse MCP provides end-to-end observability to debug AI systems by centralizing traces, observations, sessions, exceptions, and prompts for rapid diagnosis and remediation.

Core Features & Use Cases

  • Centralized observability: debug traces, find exceptions, analyze sessions, and manage prompts through Langfuse MCP.
  • MCP setup and configuration: guided setup steps for Claude Code and Codex CLI workflows, including read-only mode for safe experimentation.
  • Dataset and prompt management: list, create, and upsert datasets and prompts, with versioned prompts and labels.

Quick Start

Get started with Langfuse MCP by obtaining API keys and installing MCP via uvx. Use claude mcp add or codex mcp add to configure credentials, then verify with /mcp or codex mcp list. For a safe, read-only session, enable LANGFUSE_MCP_READ_ONLY.

Frequently Asked Questions about langfuse

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I debug AI traces and exceptions in Claude Code workflows?▼

Debug AI traces and exceptions in Claude Code workflows by centralizing observations and sessions through the Langfuse MCP, which requires API keys and host URL configuration for rapid diagnosis.

What is the best way to manage and version prompts for AI observability?▼

Manage and version prompts for AI observability by listing, creating, and upserting datasets with versioned labels through the Langfuse MCP interface to evaluate and track prompt changes.

How do I configure MCP for Codex CLI using uvx?▼

Configure MCP for Codex CLI using uvx by running the installation command with Python 3.11, then use codex mcp add to set credentials and verify the setup with codex mcp list.

Can I use a read-only mode to safely debug AI sessions without writing data?▼

Yes, you can safely debug AI sessions without writing data by enabling the LANGFUSE_MCP_READ_ONLY environment variable to disable write operations during experimentation.

Do I need Langfuse API keys to investigate errors with the MCP server?▼

Yes, you need Langfuse API keys and a host URL to investigate errors, as these credentials authenticate the MCP server connection for accessing traces and observations.

Related Skills

Explore More Software Engineering Skills→
renderincrenderinc

render-mcp

Configure and connect the Render MCP server for Cursor, Claude Code, and Codex.

Official
Intermediate
informatico-madridinformatico-madrid
1

phoenix-observability

Trace, evaluate, and monitor LLM applications with OpenTelemetry.

Community
Advanced
tavvatavva

langfuse-cli

Query Langfuse traces, sessions, observations, scores, metrics, and datasets via CLI.

Community
Advanced
jurislmjurislm

langfuse

Manage Langfuse prompts, traces, observations, and scores via MCP.

Official
Intermediate