By QodoTel Aviv, Israel1 product across 1 category
Strong fit. Writes unit tests against existing code and suggests the cases you did not think of, inside the editor.
Free plan
Yes
Free tier for individuals
Month to month
$30/mo
Pro Team, sold to teams rather than to one person
Billed yearly
—
No yearly price published
Watch out: Generated tests can assert current behaviour rather than correct behaviour, which passes and proves nothing. One plan sold in credit packs rather than a ladder: the price is the pack you pick, it is designed for up to 30 users, and overage is capped at a figure you set.
By AnthropicSan Francisco, United States3 products across 7 categories
Built for this. Reads long files and whole error traces without losing the thread, which makes it good at explaining unfamiliar code and debugging in a chat window.
Free plan
Yes
Free tier with daily caps
Month to month
$20/mo
Pro
Billed yearly
$17/mo
$200 a year, Pro plan −15%
Watch out: The chat window has no access to your project. For multi-file work you need the agent below or a dedicated editor.
By CognitionSan Francisco, United States1 product across 1 category
Strong fit. Autonomous agent that takes a ticket and works it to completion. There is a free tier, with paid plans above it.
Free plan
Yes
Free plan at $0
Month to month
$20/mo
Pro
Billed yearly
—
No yearly price published
Watch out: Autonomy means less visibility into what changed — review discipline matters more, not less. Larger organisations may still be routed through sales rather than self-serve.
By GitHubSan Francisco, United States2 products across 1 category
Built for this. The default, and the one already wired into the editor and the pull request your team uses. A free tier now covers light use.
Free plan
Yes
2,000 completions a month
Month to month
$10/mo
Pro
Billed yearly
—
No yearly price published
Watch out: Suggestions are strongest on common patterns and weakest exactly where your codebase is unusual. Pro and above bundle a monthly credit allowance, so premium-model work is metered on top of the seat.
By Augment ComputingPalo Alto, United States1 product across 1 category
Strong fit. Built for codebases large enough that context is the whole problem, and it indexes the repository rather than reading files on demand.
Free plan
No
Paid from the start
Month to month
$20/mo
Standard
Billed yearly
—
No yearly price published
Watch out: Individual plans were retired in 2026, so the entry price is a team one — flat, not per seat, which makes it unusually cheap per head on a large team. What you buy is a usage allowance rather than seats: $20 buys $20 of it, and anything beyond is topped up pay as you go.
Two things separate these and both are on every card.
Fit
Whether a tool is built for this job, a strong fit for it, or
able to do it without that being its focus.
Free plan
Whether there is a tier you can keep using rather than a
trial that expires.
The order
Fit first, then a free plan, then alphabetical. There is
no score, no weighting we decline to publish, and nothing on this site a company can buy.
Questions
What is the best ai for code review?
There is no single answer, which is why this page lists 8 rather than one: CodeRabbit, Qodo, Cursor, Claude Code, Claude, Devin, GitHub Copilot, Augment Code. 5 of them are built for this kind of work rather than able to do it. Start with a free one.
Which of these have a free plan?
6 of the 8: CodeRabbit, Qodo, Cursor, Claude, Devin, GitHub Copilot. Free means a real free tier you can keep using, not a trial that expires.
How was this list chosen?
By hand, from the 750 products this site tracks, and it is checked against the catalogue on every build — a product renamed or removed breaks the build rather than quietly disappearing from here. Nothing on this site can be paid for: no listing, no position, no wording.
