Crit
Crit is an open-source review surface for AI agent plans and code diffs: point at the line, tell the agent.
If reviewing agent output is a daily habit, Crit removes the worst part of that loop: scrolling diffs in a terminal. The file-based protocol means it plugs into whatever agent you already run — one command, no account, no integration project. Try Story mode first; a thematic view of a large diff beats file order almost every time.
Verified 23h ago · liveness 72/100 · cite: rightaichoice.com/tools/crit
- Solo developers reviewing Claude Code, Cursor or Copilot output every day
- Engineers who lose details scrolling terminal diffs and want a PR-style surface
- Reviewing long agent-written plans and spec folders with per-section comments
- Checking agent-built frontends on localhost or generated static HTML artifacts
- Non-developers or no-code users who want a simplified review tool
- Teams that need user management, roles or permission controls
- Developers who prefer terminal-only workflows and avoid browser interfaces
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Crit if you need role-based permissions, user management, or a shared review system of record — it is a local, single-user review surface without an audit layer.
Crit is an open-source, self-hosted CLI released at $0/mo, so its cost is infrastructure and time rather than a subscription: you pay for the machine that runs it and the setup step a fresh install needs (Homebrew tap, `go install`, Nix, or a pre-built Windows binary). Against a hosted code-review platform you are trading per-seat licence fees for the absence of roles, permissions, and audit trails. Story mode is the one usage-based cost to watch, since it is documented to carry an LLM token
In short
Crit — Crit is an open-source review surface for AI agent plans and code diffs: point at the line, tell the agent. Best for Solo developers reviewing Claude Code, Cursor or Copilot output every day, Engineers who lose details scrolling terminal diffs and want a PR-style surface, Reviewing long agent-written plans and spec folders with per-section comments. Free to use.
What's new in Crit
Checked 7 days agoAcross the latest 4 updates: 2 feature updates, 1 launch and 1 changelog entry.
v0.22.0: rendered Before/After markdown diffs and crit-tui for Claude Code
v0.22.0 adds a Document | Rendered | Diff toggle so markdown files in diff mode show old and new versions split or unified, and introduces crit-tui 0.1.0, an early Claude Code review pane requiring Claude Code 2.1.287+.
crit-tui pane sends review comments to Claude Code in one prompt
crit-tui opens a pane with the conversation, the git diff or a file; you select text, press c to comment, and s sends all new comments to Claude in one prompt. Comments route through the crit CLI, so the browser sees the same review.
crit plan --no-wait saves a plan version without a review
A new --no-wait flag on crit plan saves a plan version without triggering a review, and crit comment --json now accepts quote and quote_offset for programmatic commenting.
v0.21.1: daemon approve fixes and OpenCode v2 plugin migration
Crit v0.21.1 stops the daemon over HTTP on approve and crit stop, migrates unedited OpenCode v1 plugins to v2, and cleans up review-page comment layout, including scrolling document comments within the Pierre list.
What people actually say about Crit — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
97 mentions across 5 sources (Hacker News, YouTube, Bluesky, Stack Overflow, Lemmy) · researched Jul 6, 2026.
Average across the 5 sources that answered — each source counts once, not each post.
- +Local-first, no data leaves your machine unless shared.
- +Inline comments on plans, code diffs, live apps, and static HTML.
- +Four review modes cover plans, branches, live apps, and HTML artifacts.
- +Active development with frequent updates and responsive maintainer.
- +Git diff and round-by-round diff views for iterative reviews.
- −Community feedback is extremely sparse beyond a few HN threads.
- −No evidence of enterprise reliability or large-scale use.
- −Browser-based workflow may not suit terminal/IDE purists.
- −Little to no discussion on Reddit, Product Hunt, or App Store.
- −Unclear how well integrations with agents beyond Claude Code work.
- • No hidden costs — entirely free and open-source
Viability Score
How well maintained and how widely used is Crit? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →Key Features
- Comment on individual lines of plans, markdown docs and code diffs
- Rendered markdown review for agent-written plans and spec folders
- Syntax-highlighted branch and PR diffs with automatic change detection
- Story mode: reorganize a diff into thematic chapters, important changes first
- Pierre diff renderer with split and unified layouts, context expansion, lazy loading (v0.21.0)
- Rendered Before/After markdown view with Document | Rendered | Diff toggle (v0.22.0)
- Live app proxy with DOM-element commenting and automatic reload
- Static HTML iframe preview with sibling assets served, no dev server needed
- crit-tui 0.1.0 review pane for Claude Code: press c to comment, s to send in one prompt
- crit plan --no-wait to save a plan version without triggering a review
- crit comment --json with quote and quote_offset for programmatic commenting
- crit resume lists stored reviews in ~/.crit/reviews and reconnects to one
- Stacked PR support for Git, Jujutsu and Sapling
- GitLab merge request review using the same pull/push flow as GitHub
- File-based review protocol that works with any file-reading agent
About Crit
Crit is a local, open-source browser review surface for AI coding agent output. Your agent rewrites 14 files, drafts a 300-line plan, or spins up a frontend on localhost — Crit opens that work in a browser where you click the exact line, section, or DOM element that's wrong and drop a comment the agent then acts on. It's aimed at engineers running Claude Code, Cursor, Copilot, Codex, or similar agents daily who are tired of re-reading terminal diffs. Five modes match the artifact type: rendered markdown for plans and spec folders, syntax-highlighted branch and PR diffs, Story mode (thematic chapters, important changes first), a live-app proxy for localhost frontends with DOM-level commenting and auto-reload, and a static HTML preview that renders generated artifacts in an iframe with sibling assets served — no dev server needed. The protocol is the differentiator. Crit writes a structured review file, and any agent that can read a file and run a command can consume it — no API to integrate, no SDK, no hosted account. One-command installs cover Claude Code, Cursor, GitHub Copilot, Codex, OpenCode, Gemini CLI, Qwen Code, Hermes, Pi, Grok, Amp, Windsurf, Cline and Aider; GitLab merge requests flow through the same pull/push as GitHub, and stacked PRs work across Git, Jujutsu and Sapling. Recent releases keep tightening the loop. v0.21.0 swapped the flat-list renderer for Pierre (split/unified layouts, context expansion, lazy loading, Shiki themes). v0.22.0 (October 7, 2026) added a Document | Rendered | Diff toggle for rendered Before/After markdown, crit-tui 0.1.0 — a Claude Code pane requiring 2.1.287+ where you press c to comment and s to send everything in one prompt — plus a crit plan --no-wait flag. It's a solo-engineer tool from Tomasz Tomczyk that ships weekly, not a team system with roles and audit trails.
Behind the Verdict
We reach for Crit when the agent's output has outgrown the terminal — a 600-line plan, a 14-file branch, a frontend sitting on localhost. Line-precise commenting on rendered markdown is where it earns its keep; marking up a spec folder section by section beats re-prompting from memory every time. The reason Crit works across agents is the protocol, not a plugin ecosystem. Because the review is just a structured file the agent can read, adding a new tool doesn't require a new integration project. That's the design decision to weigh: no SDK to learn, but also nothing hosted to share with a teammate. Skip it if your team standardizes on GitHub PR review with roles and audit trails, or if you genuinely prefer staying terminal-only. Crit is local by design — the server binds to 127.0.0.1, zero telemetry, and nothing leaves the machine unless you click Share, which uploads to crit.md or a self-hosted crit-web. Anyone needing non-developer reviewers or permission controls will find it too thin. Where it bites: crit-tui is early (0.1.0) and requires Claude Code 2.1.287+, so treat the in-terminal pane as a bonus rather than the main path. The browser workflow on huge diffs has gotten faster with Pierre's lazy loading and context expansion, but the real unlock is Story mode — organizing a sprawling diff into thematic chapters so the important changes surface before you dig into lines. Closest alternative is leaning on a /review slash command or asking the agent to review itself. Crit's answer is fair: slash commands dump prose once, while Crit keeps comments attached to lines across rounds, shows round-to-round diffs, and leaves unresolved threads unresolved until you say so. If you want the review to persist rather than be re-litigated every prompt, that difference is the
Researching Crit? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Crit actually fits — and what changes day-one when you adopt it.
Your agent rewrites 14 files in a branch. You run Crit, the browser opens the diff with syntax highlighting and lazy loading, you click line 47 on one file and type 'this drops the refresh token'.
Outcome: The comment lands in Crit's structured review file, Claude Code picks it up and fixes that line, and the next round shows only what changed against your previous comments.
Your agent drafted a 300-line implementation plan. Crit renders the markdown, you comment on the two sections that are wrong rather than the whole document, and use `crit plan --no-wait` to save a version without triggering a review.
Outcome: You get per-section feedback instead of a wall of markdown, and the saved plan version gives you something to diff the next draft against.
Your agent spun up a frontend on localhost. Crit proxies the page into the review surface, you click the misaligned button, pin a comment to it, and the page reloads automatically as the agent changes it.
Outcome: DOM-level feedback instead of screenshots and prose descriptions, with the review staying on your machine rather than going to a hosted service.
Use Cases
- Review a 300-line AI-generated plan by commenting on specific sections in a rendered markdown view.
- Inspect an agent's code changes across 14 files with syntax-highlighted diffs and line-level comments.
- Provide feedback on a running localhost app by clicking misaligned elements in the live proxy.
- Share an HTML artifact review with a colleague via a one-click public link.
- Iterate with your agent over multiple rounds and see exactly what changed each round.
- Reopen a review from a previous session with `crit resume` instead of re-running the diff.
- Automate agent instructions on review approval using custom finish hooks.
- Review GitLab merge requests directly from your local environment.
Limitations
- Crit is a local-first, self-hosted review tool delivered as a CLI with an optional web server (crit-web), so there is no cloud version unless you host it yourself.
- Sharing exists as one-click review links with multiple share targets across Crit deployments (v0.20.1), but there is no roles, permissions, or user-management layer on top of it.
- Agent integrations depend on each agent's own plugin or skill system — v0.21.1 had to migrate unedited OpenCode v1 plugins to v2, and crit-tui requires Claude Code 2.1.287 or later.
- Story mode is documented to carry an LLM token cost.
- A Homebrew core formula is not available — installation is via the Homebrew tap, `go install`, Nix, or a pre-built GitHub binary (amd64 and arm64), so a fresh machine needs one extra step.
- Development is weekly and fast-moving, which means the surface changes often.
as of 2026-10-11
Verification history
We have re-verified Crit 10 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 10 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Crit tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Self-hosted (CLI)
$0/mo
Ideal for
Solo developers and small teams who already run Claude Code, Cursor, Copilot, Codex or Aider daily and want agent output reviewed locally without a per-seat subscription.
What this tier adds
Starting tier and the only tier: the open-source Crit CLI at $0/mo, covering all five review modes, one-command installs for 14 agents, and local storage in ~/.crit/reviews.
Where the pricing makes sense
The company stage and team size where Crit's pricing actually pencils out — and where peers do it cheaper.
Crit is an open-source, self-hosted CLI released at $0/mo, so its cost is infrastructure and time rather than a subscription: you pay for the machine that runs it and the setup step a fresh install needs (Homebrew tap, `go install`, Nix, or a pre-built Windows binary). Against a hosted code-review platform you are trading per-seat licence fees for the absence of roles, permissions, and audit trails. Story mode is the one usage-based cost to watch, since it is documented to carry an LLM token
Setup time & first value
How long it actually takes to get something useful out of Crit — broken out by persona, not the marketing-page minute.
Installation is one binary: `brew install crit` via the Homebrew tap, `go install github.com/tomasz-tomczyk/crit/cmd/crit@latest`, `nix profile install github:tomasz-tomczyk/crit`, or the pre-built Windows binary — plan a couple of minutes since there is no Homebrew core formula. Agent wiring is then a single command per agent (`crit install cursor`, `crit install aider`, or the Claude Code
Switching to or from Crit
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From pasting diffs into a chat window: run `crit install <your-agent>` once and let Crit open the branch diff in a browser instead of copying terminal output.
- →From GitHub PR review for local-only work: use Crit's branch/PR diff mode with line comments, and push to GitHub only when you actually want the PR.
- →From GitLab merge request review: Crit handles MRs through the same pull/push flow as GitHub, so you review locally before the MR is touched.
- ↗To a hosted review platform with roles and audit trails: Crit has no permission layer, so team review governance moves to the platform while Crit stays useful for solo loops.
- ↗To pure terminal review: everything Crit stores lives in ~/.crit/reviews as files, so stopping use of the browser surface leaves no data behind you need to export.
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “Crit”, and we withheld 6: 6 could not be judged, because “Crit” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Crit.
Official links
Tools that pair well with Crit
Common stack mates teams adopt alongside Crit, with the specific reason each pairing earns its keep.
Ai Review
Open-source AI code review that runs inside your CI/CD pipeline and sends diffs straight to your chosen LLM.
Shippie
Open-source AI code review agent that posts inline GitHub pull request comments on every diff.
Continue
Open-source AI coding agent for VS Code and JetBrains — free to fork since Cursor acquired Continue in January 2026.
Featured Head-to-Head Comparisons
Crit vs Cognition Ai
For enterprise teams needing an autonomous engineer that handles multi-step coding, testing, and deployment with a productivity guarantee, Cognition AI (Devin) is the clear choice. Solo developers and small teams who want a lean, local-first review loop to iterate faster with their existing agent should adopt Crit—it's free, integrates with dozens of coding agents, and respects your data privacy.
Crit vs Gem
Gem and Crit serve completely different domains — recruiting vs. code review — making a direct comparison moot. Choose Gem if you need an all-in-one AI recruiting platform with ATS, CRM, and AI agents for sourcing, screening, and scheduling. Choose Crit if you're a developer using AI coding agents and need a local-first review tool to inspect diffs, running apps, or plans. Neither is a substitute for the other.
Crit vs Poke Interaction Co
If you want an AI assistant that lives in your messaging apps to manage emails, calendar, and health data, Poke is the clear choice — especially with its new Apple Messages integration. If you need a dedicated review interface for AI-generated code (plans, diffs, live apps) and want to iterate faster with your coding agent, Crit is purpose-built and free. They solve completely different problems; pick based on whether you need task automation or code review.
Crit vs Replit Agent
Choose Replit Agent if your goal is to build and ship full-stack apps from scratch using natural language, with minimal setup for hosting and collaboration. Pick Crit if you already work with AI coding agents (like Claude Code or Cursor) and want a dedicated, local-first review interface to catch line-level issues in diffs, live apps, or plans. They serve different phases: Agent for creation, Crit for inspection.
Alternatives to Crit
View all
Frequently Asked Questions
Categories
Used Crit? Help shape our editorial sentiment research.