The cross-repo review layer

Your review bot flags 40 things. Limn shows you the 3 that matter.

Bring your own reviewer — CodeRabbit, Greptile, Copilot, whatever you run. Limn is the calm, cross-repo layer above it: what’s stalled, whose turn it is, and which of the bot’s comments a human still needs to read. One fast timeline across every repo — and the AI spend stays yours to control.

Sign in with GitHub, or run it entirely on your machine — local mode keeps no stored credentials.

Works with
CodeRabbit
Greptile
Copilot
Qodo
Sourcery
limn · bot triage

Your review bot’s output, triaged: what a commit already addressed vs what still needs a human — and clear the stale ones in one click.

01 / Problem

AI writes the code. A bot reviews the code. Who’s reading 300 bot comments a week?

Review bots are genuinely useful — and they never stop. A busy team fields hundreds of AI review comments a week, and most go unread. The firehose buries the handful that actually needed a human, across more repos than anyone can hold in their head. The bottleneck moved from writing the change to noticing what matters.

Limn doesn’t add another bot. It sits above the ones you already run: every review thread — human or bot — becomes a triaged signal. See which of CodeRabbit’s comments a commit already addressed, which still need a look, and clear the stale ones in a click.

One calm, cross-repo layer for exactly this: human-in-the-loop triage for the high-throughput era.

GitHub is a firehose wearing a UI — slow to navigate, endless tabs, notifications from humans and bots alike. Limn pulls it all into one place and gets out of your way. And it’s fast. Genuinely, annoyingly fast.

300 signals · 3 need you

Vermilion = a human is still required. The only figure on the page, and the same sprite family the game uses.

Who it’s for

For engineering managers

  • Sprint-oriented reports on blockers, what needs attention, and where throughput is improving.
  • DORA-style flow metrics you can drill into — mirrors, not scorecards.
  • Reviewer suggestions drawn from who actually touched the changed files — requested in one click.
  • A reliable state of play, in-app and in Slack, prioritised by what’s actually waiting.

For engineers

  • Track your PRs — and every PR you participate in — across all your team’s repos.
  • Know instantly when it’s your turn, without keeping forty tabs warm.
  • One-click AI review that remembers how you review, CI-failure analysis, and fixes pushed straight to the branch. When you say so.
  • The morning “what needs me?” reconstruction — gone. It’s one feed, already sorted.
02 / Free

Most dashboards add tabs. This one closes them.

Behind the feed sits the board — every repo, every contributor, every PR, review and CI run on one interactive timeline. Duration and staleness live in the shape: a long bar with no recent markers is a stalled PR, no query required.

Pick a repo and its console pulls the same picture into focus — stats, a thread-state bar, and every open PR with its CI and approval standing. The feed, the board, the consoles, the PR detail, the write actions — all of it free, open-core, forever.

Everything in the free tier →
repos down the side, time across the top
03 / Pro

AI summaries that make sense of the week — not another bot shouting into your PRs.

Per-repo digests and sprint reports, each one chained from the last — what changed since you last looked, with clickable PR refs. Delivered to Slack on your cadence, daily or twice daily.

It’s the antidote to notification fatigue: pull, don’t push. One high-quality report instead of forty pings.

The whole intelligence layer →
Not another review bot

The review-bot aisle is full. This is the shelf above it.

Review bots comment on one PR at a time — and even the good ones still bury you: independent audits put roughly a third of bot comments as noise. Limn isn’t competing to shout louder on your diffs.

It’s cross-repo situational awareness: all high-value, pull-based information — who’s blocked, what’s stalled, which threads sit unanswered — with AI review as one input you control, not the product.

And when the AI does review, it reviews your way: every run learns from what you kept, cut and reworded last time — so the noise goes down with use, not up.

AI-written sprint summary
Pro · BYO key

Yes, your CLI can do this. In eleven steps.

You could do all of this in your IDE or CLI — clone, checkout, analyse the log, prompt the agent, apply, push. Repeat for every PR, every day. Limn makes each loop one click, including the git merge conflicts.

01CI fails. Limn pulls the failing job log and diagnoses it.
02You approve a fix run. The agent patches in an ephemeral worktree — you review the actual diff.
03The fix is pushed to the branch. PR green.
01Claude reviews the PR into structured, line-anchored findings.
02You tick the findings worth keeping and write your verdict.
03One GitHub review is posted. Yours, not the bot’s.
Walk through both flows, screen by screen →
04 / Local

Or keep it entirely on your machine.

One command. No accounts, no hosted backend, no stored credentials — it authenticates with your gh CLI, syncs to a local SQLite file, and opens straight to the Activity console.

What happens when you run it →
zsh · ~/work
$ npx pierre-review
05 / Price

Free where it matters. $15 where it counts.

Free

$0, forever

The whole dashboard — Activity feed, timeline, PR detail, write actions. Unlimited repos, local-first.

Pro

$15/month

The intelligence layer — AI summaries, team Insights, flow metrics, Slack digests, My Turn.

Stop reconstructing the day from notifications.

Sign in with GitHub and the recent timeline fills in seconds, while the full history backfills behind it.