An AI agent skill that guides you through integrating Latitude telemetry into any backend project for LLM observability.
npx skills add latitude-dev/telemetry-skill
Once installed, ask your AI agent to add Latitude telemetry to your project. The skill will determine the best approach for your stack and walk you through the implementation.
Examples:
Add Latitude telemetry to my project
I want to trace my LLM calls with Latitude
Set up observability for my OpenAI calls
When activated, the skill walks you through choosing and implementing the right telemetry approach for your stack. It covers four integration paths:
| Approach | Languages | Complexity | Observability |
|---|---|---|---|
| Gateway Mode | Any | Minimal | Full (automatic) |
| SDK Telemetry | Python, TypeScript | Low | Full (tokens, cost, latency) |
| Raw OpenTelemetry | Any (with OTel) | High | Full (tokens, cost, latency) |
| Create Log API | Any (with HTTP) | Very low | Conversation logs only |
- A Latitude account and project
- Your Latitude API key (found in Settings)
- Your project ID (the number in the project URL)
SKILL.md-- Instructions for the agentreferences/-- Detailed integration guidesgateway.md-- Latitude as LLM gatewaysdk-telemetry.md-- Python/TypeScript SDK instrumentationraw-opentelemetry.md-- Manual OTLP span exportlog-api.md-- Simple HTTP log endpointspan-attributes.md-- Full span attribute reference
MIT