Markers & Screenshots

Markers

Markers annotate the recording timeline at significant moments. They are the primary tool for targeting session_diff comparisons and scoping session_replay_context.

Placing Markers

CLI

krometrail browser mark "user submitted checkout form"
krometrail browser mark "error modal appeared"
krometrail browser mark "payment failed"

MCP: chrome_mark

{ "label": "user submitted checkout form" }

Markers are timestamped at the moment they are created. They appear in session_overview output and can be referenced by name in session_diff and session_replay_context.

Using Markers in Investigation

# Get all markers from a session
krometrail session overview <session-id>

# Diff between two marker points
krometrail session diff <session-id> \
	--from-marker "user submitted checkout form" \
	--to-marker "error modal appeared"

# Generate reproduction steps scoped to marker range
krometrail session replay-context <session-id> \
	--from-marker "page loaded" \
	--to-marker "error modal appeared"

Screenshots

Screenshots are captured automatically during recording at two triggers:

Periodic capture — a screenshot is taken at a configurable interval (default: every 5 seconds) while recording is active.

Navigation-triggered — a screenshot is taken on every page navigation (URL change).

Screenshots in Investigation

Every session_inspect response includes the nearest screenshot to the event being inspected. This shows what the UI looked like at the moment the event occurred — useful for correlating a network error with a visible UI state.

krometrail session inspect <session-id> --event-id <event-id>
# Response includes: event details + nearest screenshot

Screenshot Format

Screenshots are stored as PNG files in the session database and returned as base64-encoded data in API responses, or saved to disk when using the CLI with --save-screenshots.

Tips for Effective Marking

• Mark before and after actions — "before form submit" and "after form submit" gives session_diff a precise window
• Mark when errors appear — place a marker immediately when you see unexpected behavior in the browser
• Name markers descriptively — marker labels appear in diffs and replay contexts, so descriptive names make the output easier to read
• Mark test scenario boundaries — if testing multiple scenarios in one session, mark the start of each scenario to keep them separable