May – June 2026
screencli now tests every pull request in a real browser — and emails you the verdict.
New
AI browser tests on every PR
Install the GitHub App and each pull request gets an AI-driven browser test on its preview deploy — a pass/fail verdict and a recording, posted as a PR comment before you start reviewing.
New
Bring your own Anthropic API key
Run tests on your own Anthropic account instead of screencli credits. Your key, your usage, your bill.
New
Team workspaces
Invite teammates by email and share verdicts, run history, and settings across the whole team.
Improved
Email alerts when a run finishes
The PR author gets emailed the result the moment an expect run completes — no sitting on the PR waiting for the comment.
Also in this update
- New credits model: 1 agent step = 1 credit, with 500 credits a month on Pro, refilled automatically.
- Faster CLI agent — prompt caching and a tighter prompt trim latency on every run.
- Pro plan introduced: cloud recording, the GitHub App, and the CLI agent.
- Larger agent step budget (75, up from 50) so longer flows finish.
- New screencli-test-pipeline skill for driving the test pipeline from Claude Code.
- Hardened billing — cleaner subscription renewals, refunds, and plan downgrades.