Limn · code review on your machine

Your team writes more code than it can review.

Limn reviews every pull request against its story and your docs on your own Claude or Codex, shows each person what needs them next, and fixes what it finds. Free, on your machine.

npx limn-review
Nicholas Hilliard's self-portrait miniature, 1577

“The line without shadow showeth all to good judgment.”

Nicholas Hilliard, The Arte of Limning, c.1600
Pending · Acme paymentsyou: dana

Pull requests where a person is waiting on you.

Waiting on you

Your review
Asked 26 working hours ago · +420 −88Large
  • Priya asked for your review.
  • 2 of 3 in stack: #1189 lands first.
Reply owed
Priya replied to your thread

Works with what you have

Nothing to migrate. Limn reads each tool where it lives.

  • GitHubCode, CI, issues
  • ClaudeYour login
  • CodexYour ChatGPT plan
  • Jira · LinearStories
  • Confluence · NotionDocs
  • SonarCloudQuality gate
  • SlackTeam digest
  • CodeRabbit · Copilot · Cursor+14 review bots

Results

From 7 h to 42 min

Half of pull requests merged within this much working time.

From 20 to 44

Pull requests merged per week.

$0

Extra spend. It ran on the AI plans the team already had.

A nine-repository delivery team, September vs October 2026 after Limn auto-review introduced. Measured by Limn's Chronology report.

Review

Check the decisions, not every line. Limn does the first read and tells you where to look.

Pending

What needs you next

Every open pull request across every repo, sorted by whose turn it is.

Pending

Pull requests where a person is waiting on you.

Waiting on you

Your review
Asked 26 working hours ago · +420 −88Large
Reply owed
Priya replied to your thread

AI Review

A review on every push

Your own agent reviews each new commit and posts it to GitHub, if you switch that on.

#612 Partial refundsPosted to GitHub
Open
MajorYou and PriyaRefundPolicy.allows() never reads paid_at.
MinorYouRetried refund sends a second email.
NitYouUnused import in policy.py.

Story and docs check

Built what was asked

Each change is checked against its ticket's criteria and your own docs. Every finding cites its source.

Fix

Findings, red CI and merge conflicts, fixed on your branch. Nothing is pushed until you say so.

AI Fix

Fix what the review found

Pick the findings. Limn writes the change and shows you the diff.

AI Fix · #612

CI diagnosis

Why the build is red

Limn reads the failing log and says whether it is your code or a flaky test.

test-api · failedFlaky
ConnectionResetError: [Errno 104]
  at tests/test_webhooks.py:88
Same check passed on this commit at 10:42.

Not your change: the test lost its network connection.

Conflicts

Resolve conflicts in the app

Both sides in one view, and you pick. Or AI Fix merges in its sandbox, and you check the merge before Push.

#701 · 3 conflicts
1 left
Appliedledger/post.py · took theirs
Both appliedledger/models.py · both
Needs a decisionapi/routes.py · lines 40–52

Ship

Know what can land, land it when it's ready, and keep main green.

Merge verdict

What is blocking it

One answer to "can this land?", with the reasons in order.

#1190 Settings split
Blocked2 of 3 in stackLarge
1Needs 1 more approval
2Branch is behind main

Merge when ready

Lands on its own

Arm it once. Limn merges when checks pass, and stops if anyone pushes.

Merge when ready
#476 Ledger index
Armed
Waiting for 1 check · then merges
#482 Retry budget
2 approvals · 1 check running

Trunk

Is main green?

The state of each main branch, and the pull request that broke it.

Main branches
Green
Rede2e-checkout since #1184 merged · 35 min
Green

See

Where the work is, where it waits, and what your review bots are worth.

Chronology

Where every hour goes

Each open hour belongs to someone: waiting for review, with the author, or waiting to land.

Bots

What your bots are worth

Each review bot's comments scored for how serious they are, and what they cost.

Review bots
BotCommentsHighActed on
41238%
29022%
11871%

Why Limn

Built to fit how your team already works.

On your machine

Your code and data stay on your disk. No hosted service, no sign-in.

Nothing to migrate

Uses your gh login, your CI and your tracker as they are.

Your AI plan

Runs on Claude Code or Codex. Set a weekly share and it stops there.

Set a budget →

You stay in charge

Nothing is posted or pushed until you press the button, or switch it on.

Every feature, free

Start on your own today. Bring the team when it works for you.

Individuals and teams up to 3

Free

  • Every feature: review, fix, ship, reports
  • Runs on your own AI plan
  • No account, no card

More than 3 people

A licence

  • The same app, the same features
  • One licence for the organisation
  • We answer every message

Ask for a licence

Questions

Does my code leave my machine?

Only to your own AI provider, as when you use it yourself.

Can I keep my review bots?

Yes. Limn reads them, scores them and shows what they cost.

Do I need to install Claude or Codex?

No. Limn sets them up. You only sign in.

What does it cost?

Nothing from us. Reviews run on your own Claude Code or Codex, and Limn shows what each run used.

Will it post to GitHub for me?

Yes, once you switch on auto-posting for a workspace. It posts one review per pull request.

GitHub Enterprise?

Enterprise Cloud on github.com, yes. Self-hosted Enterprise Server, not yet.

From open pull request to merged, faster.

npx limn-review

One minute to run. Your first review a few minutes later.

Install and run →