Replace code reviews with verified intentAviator Verify
Aviator

AI made code easier to write, not easier to ship.

Aviator gets each PR to the right reviewer, runs deterministic verification against what was actually asked for, and merges and deploys it without breaking main, at the volume your agents produce.

acme/billing #4823 · mbauer
Rate-limit the /webhook endpoint to prevent retry storms
Verify Results
6/7
Acceptance Criteria
1Requests above 100/min per subscriber return 429
2429 responses include a Retry-After header
3Queued retries back off exponentially, up to 5 attempts
Invariants
1Network calls set a timeout
Why it failed: httpx.post in webhooks/retry.py:88 is called without a timeout.View invariant
2No hardcoded secrets
3Structured logs on exceptions
4Migrations are backward-compatible

From startups to Fortune 500s, the world's most effective engineering teams ship with Aviator.

MetaNotionSlackFigmaSquareDoorDashCoda
SecureframeVerkadaWorkOSCyeraHeadwayAnrokPort

Review

Review the decisions, not every line. Aviator verifies each PR deterministically against its acceptance criteria and your team's invariants, so reviewers spend their time on intent and trade-offs.

Inbox

Every PR waiting on you, across repos, sorted by what to do next: quick fixes, PRs that block others, and re-reviews after new commits.

How Inbox works
Requested review4
TitleUpdated
r
Share session state across billing middleware
acme/billing #4821New changes
2h
t
Migrate billing service to the v3 Stripe SDK
acme/billing #4815
5h
k
Fix race in retry handler when token refresh fails
acme/platform #4798Conflict
1d
j
Tighten generics in useForm hooks
acme/web #4760
4d

Team Reviews

Assigns reviewers by code ownership, expertise, and current load, and tracks time to first review against your team's SLO.

How Team Reviews works
OpenRefactor billing session middleware+247 −98
Suggested reviewers, ranked by expertise in the changed files and current review load
r
rsotoRecommended
Code owner of billing/session
Request
t
tnakamura
Reviewed 14 PRs in billing/ this quarter
Request
k
kfinch
Platform on call this week
Request
Team SLO first review in 4hLast 30 days 94% inside SLO

Verify

Checks every PR deterministically against defined acceptance criteria and team rules, with auditable evidence. Captures user decisions and reviewer arguments to build team knowledge.

How Verify works
Rate-limit the /webhook endpoint #4823Approve
Decisions1
One rate limit applies to every planOpen question

Aviator agent: Make it per-plan now.

m

mbauer's agent: Follow-up. No plan sends 20 a second today.

Acceptance criteria33/3
Invariants32/3
I1Network calls set a timeout
webhooks/retry.py:88httpx.post has no timeout
I2No hardcoded secrets
I3Structured logs on exceptions
Evidence3

Ship

Keep main green at any merge volume, and always know which environment each change has reached.

MergeQueue

Parallel queues by affected target, batching, and automatic handling of flaky tests. Built for monorepos merging 1,000+ PRs a day.

How MergeQueue works
Queueacme/monorepo · main
QueuePendingOpenClosedMerged

Verifying with Draft PRs (2 batches)

Draft PR #9812Top5/7
#4790rCache plan lookups in the billing API
#4793kRetry flaky upload in the ingest worker
Draft PR #98132/7
#4801jTighten generics in useForm hooks

Waiting for the next batch (4 PRs)

Batch size 4Next batch in 3 minutes

Releases

Cut releases, deploy, cherry-pick, and roll back from one dashboard that tracks every PR across every environment.

How Releases works
releases > api-server
Release Dashboard
staging
v2.15.0-rc218m ago
production
v2.14.32d ago
v2.15.08f3c2a1staging: rc2Deploy production
v2.14.341be09dproduction: v2.14.3Cherry-picked #4798

Start with the free tools, bring the team when you're ready

Free for developers

Inbox
Every PR waiting on you, across repos and teams.
Stacked PRs CLI
Open source. Split large changes into PRs people can review.
Chrome extension
Aviator's PR workflows inline on github.com.
Slack notifications
Per-channel rules, so PR pings stop drowning out everything else.
Get started free

For engineering teams

Team Reviews
Reviewer routing, review SLOs, and team dashboards.
Verify
Deterministic checks against acceptance criteria and team rules, and a decision review framework to build team knowledge.
MergeQueue
Keeps main green at any merge volume.
Releases
Deployments, rollbacks, and cherry-picks in one place.
Built on engineering fundamentals
Trunk-based development, small PRs, monorepos, and continuous deployment worked before AI. We make them hold up at AI volume. How we think about it
Works with the CI you have
Aviator reads the checks your CI already reports, so there's nothing to migrate and nothing to lock you in.
On-prem or hosted
Run Aviator in your own cloud, or use our hosted version, which is SOC 2 Type II audited.
Support from people you know
Customers get a dedicated customer success contact and a shared Slack channel for real-time help.
Proven at enterprise scale
More than 1,000 teams, from 10-person startups to Fortune 500 monorepos.

What customers say

Notion
“Aviator's merge queue finally fits how we ship in a monorepo: it supports stacked PRs and parallelizes queues by affected target instead of forcing everything through one serial bottleneck. As our AI tooling increased throughput, Aviator made merging fast and reliable again at our volume — we've cut time-to-merge for p50 of all PRs by 40%.”

Erdem Alparslan

Head of Developer Experience

40%

faster p50 time to merge

DoorDash
“Aviator has enabled us to do a lot more things, like automated cut releases and deployments. I don't think we could have done that without Aviator. We've seen Aviator catch a lot of the mainline issues ahead of time, preventing outages.”

Jamil Seaidoun

Staff Engineer

Reforge logo
“Our team loves MergeQueue. We were getting a lot of complaints from engineers that were blocked on their PRs waiting to merge. We tested and implemented Aviator MergeQueue. Engineers are now happy, and we're merging at a speed that would not have been possible without MergeQueue.”
Adi Ranjan, Engineering Manager
Shippabo logo

Shippabo

“Aviator was an immediate, and I would say massive, time saver. Definitely a sanity saver for us. One of the reasons we weren't doing releases more often than every couple of days was this contention point of managing PRs was tough to deal with. Since then, we deploy all the time.”
Ben Tea, Staff Software Engineer
Prodigy logo

Prodigy

“My favorite feature of Aviator is that you don't see Aviator. You add the label in GitHub, and you go do something else.”
Chris Light, Principal Software Engineer
Secureframe logo

Secureframe

“Aviator has saved our team a lot of headaches associated with merge processes. I no longer have to worry about insidious semantic conflicts that break our builds and drive development to a halt. Managing our main branch has become frustration-free; our builds are always deploy-ready.”
Chintan Parikh, Head of Compliance Engineering
Butternut Box logo

Butternut Box

“Aviator has massively relieved developer frustration in terms of having to constantly monitor PRs. It has given our team the confidence to “set and forget” many previously manual workflows — fantastic!”
Niall Paterson, Director of Engineering
Verkada logo
“Aviator Releases easily saves us 10-15 minutes when responding to urgent incidents compared to our previous tooling.”
OJ Kwon, Engineering Lead
Amplitude logo

Amplitude

“Teams that are working in a monorepo where 100+ engineers merge all at the same time and things are highly coupled with each other… if those types of teams start seeing that their main branch is breaking once a day or more, I think that's probably the time to consider using a tool like Aviator.”
Mae Capozzi, Senior Software Engineer

The Hangar, a community for developer-experience leaders

It's where DevEx leads from Slack, LinkedIn, DoorDash, Figma, Google, Docker, and others trade hard-earned wisdom on building the future of software engineering.

Join the community

The Hangar DX Podcast

Stories from engineering teams on developer experience and shipping at scale.

Featured episode

Engineering Discipline in the AI Era with Dave Farley

The author of Continuous Delivery on why AI is a bigger shift than Agile, how it reshapes programming into specification and verification, and why engineering discipline matters more than ever.

45 min

From open PR to production, faster

We'll walk through how your team reviews, merges, and ships, and show you where the time goes. Or start with the free tools on your own.