Changelog

• The Playwright and Appium MCP presets are now enabled by default, so browser automation tools are available to your AI agent without opting in first.
• Failure screenshots now render inline in the AI conversation, instead of requiring a separate click-through.

The Tests view in Studio is rebuilt around a full-width list with prefix search and grouping.

Runs, repos, test cases, audit logs, AI conversations, eval runs, alerts, members, API keys, and MCP tokens all switch from paged loading to infinite scroll, so large orgs and repos load noticeably faster — the full table no longer has to download before you see anything. The Repositories list also gains sorting by last run, name, pass rate, and recently added.

A new Settings home adds a readiness checklist and a jump grid to every section:

• ⌘K now searches inside settings too — around 30 individual fields, not just page names — showing the current value and deep-linking straight to it.
• Simple toggles save instantly with a 5-second "Saved · Undo"; multi-field forms collect changes into one sticky Save bar, so it's always clear whether something saved.
• A new Preferences page covers theme, table row density, default start page, timezone, and date format.

Scheduled and one-off automation jobs are now unified, with real wall-clock schedules — time of day plus which weekdays — instead of a bare interval. Failure tracking shows streaks, finished runs trigger a desktop notification, and you can edit, retry, cancel, or undo a run directly from Automations.

Studio and the CLI/TUI get a real Git surface — merge, rebase, resolve conflicts, and restore points with undo. Run it from Studio's UI or from the terminal with /git ...; AI agents can propose changes through the same surface but are never allowed to force-push, reset, or discard on their own.

A run now opens on its failures by default, grouped by error signature and tagged "new" or "since #N" against the previous run:

• Compare any two runs to see what's newly failing, newly passing, still failing, or stopped running.
• Pull up a test's history across every run it's been part of, filter by branch, and group by error or suite.
• Ask AI opens pre-filled with the failure you're looking at — no re-explaining what broke.
• The Test Runs list itself gets resizable, reorderable columns, an activity calendar, and the same search and filter bar as the rest of the product.

Repo pages are rebuilt around one unified tab row — Overview, Test Cases, Test Runs, Navigations, Settings:

• Overview is the new default landing tab: an AI-generated read of your latest run (likely defect, severity, root cause, suggested fix) plus DOM failure hotspots showing which pages and elements fail most often.
• Test-case status counts — Passed, Failed, Flaky, Skipped, Timed out — sit in the page header.
• Test Cases gets List, Compact, and Suites views, with a field picker that remembers your choice.