> ## Documentation Index
> Fetch the complete documentation index at: https://docs.squasher.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Session replay

> Watch real user sessions with correlated errors, logs, and traces in one workflow.

Session replay lets you move from "an error happened" to "this is exactly what the user saw" without leaving Squasher.

## What you get

* Playback controls up to `4x`, including fast review at `2x`.
* Automatic pause skipping so idle time does not waste review time.
* Correlated errors, logs, and traces alongside the active replay.
* Built-in sampling so you can control session volume.
* Privacy controls for masking text and blocking sensitive DOM regions.

## How it works

Enable replay in the [Browser SDK](/sdks/browser) and Squasher will batch rrweb events in the background, persist replay segments, and attach session identity to frontend telemetry when available.

Once a session is recorded, the replay detail view shows:

* page transitions
* replay-linked error markers
* AI tags, highlight state, and enrichment status for sessions worth reviewing
* network waterfall timing
* backend logs tied to the same session
* distributed traces carrying the same `session_id`

The session list and replay viewer stay in the same Explore workspace. The selected session is stored
in the URL, so reload, shared links, and browser history return to the same replay. Use the related
telemetry actions to open scoped logs or traces with the same time range and environment.

## Good defaults

* Start with a low sampling rate in production.
* Mask all text or block known sensitive selectors before recording customer traffic.
* Use replay together with browser errors and traces so the session timeline has useful backend context.

## Retention

Replay retention follows your plan's replay-specific retention window. Expired or deleted sessions keep a readable metadata state when available, but raw segment payload reads are blocked after retention.

## Agent handoff

For agents, start with compact replay search and inspect responses instead of fetching raw segment payloads:

```text theme={null}
Investigate replay sessions for project <project_id> between <from> and <to>. Search for sessions with errors or failed network requests, inspect the strongest candidate, and ask before fetching raw replay segment events.
```

Use the [Replays API](/api-reference/replays), `squasher replays search`, `squasher replays inspect`, or MCP `search("replays.search")` and `search("replays.inspect")`.