Get started – free
  • Home
  • Self-driving product
  • Context warehouse
  • Pricing
  • Docs
  • Demo
  • Talk to a human
  • Why PostHog?
  • Changelog
  • Company handbook
  • Store
  • Careers
  • Trash
  • Home
  • Self-driving product
  • Context warehouse
  • Pricing
  • Docs
  • Demo
  • Talk to a human
  • Why PostHog?
  • Changelog
  • Company handbook
  • Store
  • Careers
  • Trash
AI Observability
Overview
Getting started
Start here
Installation
Privacy mode
Concepts
Basics
Generations
Traces
Spans
Sessions
ErrorsBeta
Tools
Embeddings
Evaluations
Prompt management
Guides
Calculating LLM costs
Analyze LLM performance
Tracking custom properties
Link Session Replay
Link Error Tracking
Collect user feedback
PostHog AI
Analyze traces with PostHog AI
Query traces with MCP
Playground
Trace summarizationAlpha
ClustersBeta
SentimentBeta
Resources
Custom parsersBeta
Demos & examples
Data retention
Troubleshooting
Changelog
More tutorials
Third-party integrations

Getting started

AI Observability

Capture traces, generations, and spans from your AI and LLM products.

Start capturing LLM data

Resources

Real-world use cases to get you started

  • Guide

    How to set up OpenAI observability

    Track API usage, cost, and latency.

  • Guide

    How to set up AI Observability for Cohere

    Track your Cohere usage, costs, and latency.

  • Guide

    How to set up AI Observability for Anthropic's Claude

    Monitor costs per model, user, and more.

Explore guides

Community questions