OpenSkillIndex← back to search
claude-codedevelopment

Babysitter langfuse-integration

LangFuse LLM observability integration for tracing, analytics, and cost tracking

install
source · Clone the upstream repo
git clone https://github.com/a5c-ai/babysitter
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/a5c-ai/babysitter "$T" && mkdir -p ~/.claude/skills && cp -r "$T/library/specializations/ai-agents-conversational/skills/langfuse-integration" ~/.claude/skills/a5c-ai-babysitter-langfuse-integration && rm -rf "$T"
manifest: library/specializations/ai-agents-conversational/skills/langfuse-integration/SKILL.md
tags
#llm-observability#cost-tracking#analytics#tracing#langchain#langfuse
source content

LangFuse Integration Skill

Capabilities

  • Set up LangFuse tracing for LLM calls
  • Configure cost tracking and analytics
  • Implement prompt management
  • Set up evaluation datasets
  • Design custom trace metadata
  • Create dashboards and alerts

Target Processes

  • llm-observability-monitoring
  • cost-optimization-llm

Implementation Details

Core Features

  1. Tracing: Track LLM calls, chains, and agents
  2. Prompts: Version and manage prompts
  3. Analytics: Usage, latency, cost metrics
  4. Datasets: Evaluation and testing data
  5. Scores: Track output quality

Integration Methods

  • LangChain callback handler
  • Direct SDK integration
  • OpenAI drop-in replacement
  • Decorator-based tracing

Configuration Options

  • Public/secret keys
  • Host URL (cloud or self-hosted)
  • Sampling rate
  • Metadata configuration
  • User tracking

Best Practices

  • Consistent trace naming
  • Meaningful metadata
  • Regular prompt versioning
  • Set up alerting

Dependencies

  • langfuse
  • langchain (for callback integration)