Crit
Verified Oct 2, 2026 ActiveOpen-source local review tool where you comment on an AI agent's plans, diffs and running app in the browser.
Listed on EffectHub ·
Pricing
- Pricing model
- 100% Free
Free and MIT-licensed. The CLI runs locally; the crit.md site offers optional public share links and GitHub sign-in at no charge, and the share server can be self-hosted. No paid plan is listed.
Use tool ↗About Crit
Crit opens what a coding agent produced (a markdown plan, a branch or PR diff, a static HTML page, or a live dev server) in a browser, where you leave PR-style inline comments on any line or element. It writes those comments to a structured review file that the agent reads, so any agent that can read a file and run a command can take part. Each round shows what changed since the last one, and a story mode groups a large diff into themes. Install via Homebrew, Go or Nix; an optional web server can be self-hosted with Docker for shared review links. It is a one-person MIT-licensed project.
- API
- No
- Last verified
- Oct 2, 2026
Verified sources
- Review modes, agent integrations and local-first positioningcrit.md ↗
- Terms state Crit is MIT-licensed and operated by Tomasz Tomczyk as an independent projectcrit.md ↗
- MIT-licensed repository, created February 2026github.com ↗
- Self-hosting the share server with Docker Compose and Postgrescrit.md ↗
Is Crit legit?
We don’t label tools safe or unsafe. Below are the records we checked ourselves, each with the date we read it.
- Last checked
- Oct 2, 2026The site answered
- On the market since
- 2026The vendor's own material, read by us
- Operated by
- Tomasz Tomczyk, United KingdomOur research, card verified Oct 2, 2026
Pros & cons of Crit
Crit fixes the awkward part of working with coding agents: reading a 14-file diff or a long plan in a terminal and then typing a vague correction. It suits developers who already run Claude Code, Cursor, Codex or a similar agent and want PR-style comments sent back as structured feedback. It is a young solo project, so expect fast change and thin support.
Pros
- Works with any agent that can read a file, with install helpers for about a dozen of them
- Round-by-round diffs show exactly what the agent changed after each comment
- Covers plans, diffs, static HTML and live running apps, not only code
- Local-first and MIT-licensed, with a Docker-based self-host option for sharing
Cons
- Created in February 2026 and maintained by one person
- It collects and relays human comments; it does not review the code itself
- Shared review links are secret URLs, so access control is only as good as the link
Best for
- Developers reviewing agent-written plans and diffs before accepting them
- People who use several different coding agents and want one review surface
- Teams that want shared review links without sending data to a third party
Not for
- Teams that need a hosted, supported review platform with SLAs
- Anyone who wants an AI that reviews the code for them rather than a place to comment
Watch out: The public share feature uploads your document and comments to the crit.md server; use the self-hosted server if the content is sensitive.
Top alternatives to Crit
Crit head-to-head
Side-by-side fact sheets against the competitors our research or their makers name — pricing, platforms, API and both verdicts on one page.
Also used for
AI tools by profession
Code Review & Quality is one of the jobs on this page: