The Mesrai CLI brings the cloud review engine onto your machine. Run it against your working tree, your staged diff, a feature branch, or as part of a CI step — same multi-agent reviewer, same Mesrai Rules, same models, just before the PR ever exists.
Who should use it
- Solo developers wanting a sanity check before they push
- Teams that want one rule set shared across every laptop, IDE, and CI runner — no per-developer login
- AI coding agents (Claude Code, Cursor, Codex, Windsurf) running review-fix loops with machine-readable output
- CI/CD pipelines failing builds on any finding above a severity threshold
What a run looks like
mesrai review produces a structured report — files touched, findings grouped by severity, preview-able fix patches, and a script-friendly exit code.
Pass --prompt-only to swap the human-friendly view for a compact, structured payload an agent can parse.
What you can do
Interactive review
Step through files, preview each fix, apply or skip one by one.
Bulk auto-fix
Apply every fixable finding in one shot with --fix.
Agent integration
Pipe --prompt-only into Claude Code, Cursor, or Windsurf for a hands-off review-fix loop.
Pre-push hook
Block a push on findings above a threshold — mesrai hook install.
Fetch PR suggestions
Pull Mesrai’s suggestions for an open PR straight into your terminal.
Business validation
Check that a local diff actually satisfies a task’s acceptance criteria.
Mesrai Rules
Create, edit, and inspect the team rule set the reviewer applies.
Decision memory
Persist agent decisions as structured files versioned next to your code.
Usage limits
| Anonymous trial | Signed-in account | |
|---|---|---|
| Reviews per day | 5 | Per your plan |
| Files per review | 10 | 100 |
| Lines per file | 500 | — |
| Max diff size per file | 500 KB | 500 KB |
| Max content per file | 2 MB | 2 MB |
Trial counters reset every 24h. Check current usage with mesrai auth status.
CLI vs web dashboard
CLI and dashboard share the same engine, the same rules, and the same repo settings. Use the CLI when you want a review before a PR exists — on a working branch, a staged diff, or inside an agent loop. Use the dashboard for the post-PR experience: review thread, analytics, team config.
Rule + config edits via mesrai config repo and mesrai rules propagate to both surfaces.