The call
Start on Essentials at US$24 a developer a month billed annually, and assign seats by hand to the people who open pull requests.
Set over-limit reviews to On demand with a monthly cap. Move to Team at US$48 only for triage, unit test generation, or 8 reviews an hour.
Seats set CodeRabbit's base bill, and pushes decide how fast each developer reaches the hourly review limit. Essentials allows 5 reviews per developer an hour, and a coding agent that pushes after each fix can spend them before the hour is out. A merge then waits on the rate limit.
When you need this
- You are moving from the 14-day trial to a paid plan.
- Coding agents or bots open pull requests and push many times per change.
- Developers see review-limit messages, or finance asks what review costs per month.
Worked example: a 25-developer team sizing its bill
Seats
Twenty-five developers on Essentials billed annually cost US$600 a month. On Team the same seats cost US$1,200 (GILD calculation from CodeRabbit's plans page).
The hourly allowance
Essentials includes 5 PR reviews per developer per hour. Every push, force-push, or rebase on an open pull request spends one, so a developer whose coding agent pushes six fixes in an hour hits the limit.
Past the limit
With On demand set, extra reviews cost US$0.25 per reviewed file. Fifty extra reviews of 8 files each add 400 files, or US$100 that month (GILD calculation).
What to change first
Set auto_pause_after_reviewed_commits to 1 or 2 or turn off auto_incremental_review, as Nathan Payne did after agent edits used up his 5 reviews an hour, before paying for Team.
Evaluation criteria
Choose the plan by feature
Which features do you need in the first quarter?
Look forEssentials costs US$24 per developer per month billed annually, or US$30 monthly. It includes PR reviews, the knowledge base, linters, and usage-based reviews. Team costs US$48 or US$60 and adds triage and unit test generation, plus merge-conflict resolution. Advanced costs US$72 or US$90 and adds continuous pull request security review. Enterprise has custom pricing and adds SSO and audit logs, plus self-hosting.
Pick a seat mode
Who should hold a seat?
Look forAfter the trial, assign seats in Team Management under manual approval. Developers can request seats from pull requests. Auto-approval assigns seats to pull request authors and reassigns seats unused for 30 days. On non-Enterprise plans, it can buy prorated seats when all are taken. Exclude bots and service accounts with ignore_usernames.
Know the hourly allowance
How many reviews does each developer get?
Look forEssentials includes 5 PR reviews per developer per hour, Team 8, Advanced 10, and Enterprise 12. Open source projects get 1 to 10, depending on the project. Allowances refill on a rolling window. Free summarizes pull requests and reviews code only in the IDE and CLI.
Stop spending reviews on churn
What consumes a review?
Look forBy default every push triggers an incremental review. Force-pushes, rebases, reopened pull requests, and drafts marked ready also spend one. Batch commits, set auto_pause_after_reviewed_commits to 1 or 2, or turn off auto_incremental_review. Nathan Payne ran out of his 5 reviews an hour across nine agent-driven repositories because each round of agent changes triggered a review.
Decide the over-limit mode
What happens when a developer reaches the limit?
Look forOn Essentials, Team, and Advanced, choose Automatic, On demand, or Off for reviews beyond the included limit, and set a monthly spending cap. Usage-based reviews cost US$1 per credit, and a credit covers 4 files, so each reviewed file costs US$0.25. Over-limit pull request reviews are free during the initial trial.
Budget the on-demand products
What is billed outside the seat?
Look forCodeRabbit Security Scan and AI Deep Scan bill per scan. CodeRabbit Agent bills separately by agent minutes for cloud coding tasks, Slack requests, and automations, and is unpaid during its launch, with usage caps.
Watch usage
Where do you see who reached the limit?
Look forThe Review Usage dashboard shows review events and each developer's limit. Its Review Log shows which reviews a rate limit held back.
Questions to ask vendors
- How many people and bots open pull requests each month in the repositories we will connect?
- How many pushes does a typical pull request get, and should each one trigger a review?
- What monthly spending cap matches the value the pilot measured?
- Does Essentials cover the pilot, or do we need Team's triage and test generation?
- Who is alerted when usage-based spend nears the cap?
Red flags
- Auto-approval seat mode with service accounts that are regular user accounts: they take seats like any author, and ignore_usernames does not stop seat assignment.
- Usage-based reviews on Automatic with no spending cap.
- A required status check that blocks merges when CodeRabbit reaches a review limit, as Nathan Payne hit across nine repositories.
- Comparing list prices without counting how many reviews each pull request spends.
- Contractors who need a separate seat in every client organization, which Nathan Arthur names as the main problem with the pricing.
Products in this piece
Sources
- Official documentation
- 7
- Public source
- 3
- CodeRabbit pricing
- CodeRabbit docs: Plans and pricing
- CodeRabbit docs: Seat assignment
- CodeRabbit docs: Review rate limits
- CodeRabbit docs: Usage-based reviews
- CodeRabbit docs: Review Usage dashboard
- CodeRabbit docs: Automatic review controls
- Nathan Payne, mergepath issue #826: rebalance the review topology (Jul 30, 2026)
- island-io/mila pull request #221: replace CodeRabbit with Cursor Bugbot (Aug 31, 2026)
- Nathan Arthur: AI for Code Review: CodeRabbit (Aug 19, 2025)