# AI Observability in PostHog Web - Docs

The [PostHog web app](https://app.posthog.com) is where you actually read your AI product's traces. Open **AI Observability** in [PostHog](https://app.posthog.com) and you land on a dashboard of cost, latency, token usage, and error rate, with tabs for traces, sessions, users, tools, and errors alongside it.

From there you can open any trace to see the full conversation – every [generation](/docs/ai-observability/generations.md) and [span](/docs/ai-observability/spans.md) in order, with inputs, outputs, model, tokens, and cost on each one. Traces carry the person who triggered them, so you can jump straight to their [session replay](/docs/ai-observability/link-session-replay.md) or any [exception](/docs/ai-observability/link-error-tracking.md) the request threw.

## What you can do here

-   **[Analyze LLM performance](/docs/ai-observability/dashboard.md)** – track cost, latency, tokens, and error rate across models, features, and users.
-   **[Read traces and sessions](/docs/ai-observability/traces.md)** – follow a single request end to end, or a whole multi-visit [session](/docs/ai-observability/sessions.md).
-   **[Test prompts in the playground](/docs/ai-observability/playground.md)** – rerun a prompt against a different model or wording without redeploying.
-   **[Cluster similar traces](/docs/ai-observability/clusters.md)** – group conversations by what users were actually trying to do, so failure patterns stand out.
-   **[Check sentiment](/docs/ai-observability/sentiment.md)** – see which conversations left users frustrated, classified automatically.
-   **[Summarize a trace](/docs/ai-observability/summarization.md)** – get a written summary and flow diagram of a long agent run instead of reading every message.
-   **[Analyze traces with PostHog AI](/docs/ai-observability/analyze-traces-ai.md)** – ask PostHog AI, right here in the web app, about spend, latency, and errors across your trace data, and have it write [custom parsers](/docs/ai-observability/custom-parser.md) for trace formats it doesn't recognize. It's an interaction mode inside the web app, not a separate surface.

## Related

-   Start with the [data model](/docs/ai-observability/basics.md) if traces, spans, and generations are new to you.
-   Query the same data from your editor over [MCP](/docs/ai-observability/surfaces/mcp.md).
-   Drive the same things from your own code with the [API](/docs/ai-observability/surfaces/api.md).
-   Score your traces automatically with [evaluations](/docs/ai-evals.md).

### Community questions

Ask a question

### Was this page useful?

HelpfulCould be better