It runs your reviewers, not just one
Buddhi fans out to the AI reviewers you already use, gathers everything they say, and reconciles overlapping or contradictory comments into a single coherent set of work.
An agent that owns the pull request until it is done.
You already have AI reviewers. You are still the glue: trigger each one, wait, reconcile their contradictory comments, work out which are actually right, fix them by hand, re-trigger, repeat. Buddhi runs that whole cycle for you, unattended, and hands back only when the pull request is clean or it needs a decision from you.
What Buddhi does
Buddhi is a code-review assistant for developers. You point it at a pull request and it takes over the part of the job that is repetitive but still needs judgment.
It tests each claim instead of obeying it
Review bots are confidently wrong a lot. Buddhi treats every comment as a claim to be checked against your actual code, and applies only the fixes that survive its own verification.
It will not break your build
Buddhi runs your test suite before it pushes and stops if anything goes red. It never rewrites a test to fake a pass, and it never pushes a commit whose tests it has not seen go green.
It knows when to ask you
When a comment turns on a judgment call that the code and the docs cannot settle, Buddhi surfaces it with a diagnosis and waits, rather than guessing on your behalf.
What you get with a paid licence
Buddhi Review Pro is a licensed software package that runs locally on your own machine, on your own AI subscription. Your source code is sent to the AI providers you choose, under your own account with them, and is governed by their terms. It is never sent to us. We operate no service that receives, stores or processes your code. A purchase includes:
- The Buddhi Review Pro package, installed locally and run from your terminal.
- Hands-off automation through rebases, merge conflicts, failing tests, and the merge itself.
- Parallel fixing across independent files, so a wide pull request finishes far faster.
- Best-value routing of every fix across the AI models you already pay for.
- Batch review across many pull requests at once, unattended.
- Spend tracking so a team's review costs stay predictable.
- A licence valid for the duration of your subscription, with priority support.
- An unconditional 30-day money-back guarantee on every purchase.
Pricing
Free
$0
Open source, forkable, yours to keep.
The full quality and safety floor: multi-reviewer fan-out, verified fixes, a test gate that refuses to push a broken build, and an opt-in merge of a clean pull request. Runs on your own Claude plan.
Plus
$9 / developer / month
or $89 / developer / year, billed annually
A licensed package that runs locally, on your own AI subscription.
Everything in Free, plus hands-off autonomy through the messy real world: rebases, conflicts, failure recovery, parallel fixes, best-value model routing, batch review across many pull requests, and spend tracking.
Try Buddhi Review Pro free for 14 days. No card is required, and the trial ends automatically unless you choose a paid subscription.
Unconditional 30-day money-back guarantee.
All prices are in US dollars and exclude any sales tax or VAT, which is calculated and charged at checkout by Paddle. See our Terms of Service and Refund Policy.
Free and Commercial, side by side
| Benefit | Free skill | Commercial skill |
|---|---|---|
| Review quality | Drives your PR to genuinely clean — fans out your review bots, reconciles their comments, and applies only fixes that survive its own verification. It tests each reviewer's claim instead of trusting it. | Everything in Free, with the same uncompromising verification at every layer. |
| It won't break your build | Runs your tests before it pushes and stops if anything goes red. It never rewrites a test to fake a pass and never ships a broken build. | Adds autonomous recovery: when a fix trips a test, it diagnoses and repairs or rolls back on its own — then keeps going. |
| Automation depth | One command turns the babysat review-and-fix cycle into an unattended loop that runs to convergence. Merges a clean, up-to-date PR when you opt in. | Owns the messy real world hands-off: rebases, resolves conflicts, recovers, and merges — and fixes independent files in parallel so wide PRs finish far faster. |
| Scale across many PRs | Handles one PR at a time, start to finish. | Point it at a whole batch of PRs and walk away — it reviews, fixes, and merges them in parallel, unattended. |
| Intelligence | Right-sized, careful fixes on your own Claude plan — no new vendor, no extra keys. | Routes every fix to the best-value model across your whole bot fleet and tunes effort per task, cutting time and spend. |
| Stay in control, anywhere | Surfaces genuine judgment calls instead of guessing, and lets you answer from your terminal. | Answer from your phone in one tap, with pre-reasoned options — approve on the move. |
| Spend & scale | Runs locally on your existing subscription; stops gracefully before it ever hits a wall. | Tracks and paces spend against your team budget with live dashboards, so cost stays predictable at scale. |
| Setup & support | A guided, polished setup that gets you to your first run in minutes. | Priority support, team configuration, and policy controls. |
Where to get Buddhi
Buddhi installs as a skill for Claude Code. The free version is on GitHub under an open-source licence. The commercial licence is sold through Paddle, our authorised reseller.
- buddhi-review on GitHub — the free skill.
- buddhi on GitHub — the open kernel it is built on.
Contact
Buyer support, billing questions, and refund requests:
Email: [email protected]
Phone: +1 971 259 4681