The persistent visual QA layer for AI-shipped UI

Your AI agent has eyes now.
Give it a memory.

Agent skills can screenshot what was just built — a point-in-time check. UICanary is the persistent layer underneath: baselines across every branch, multi-viewport diffs on every PR, and a review workflow your team can trust for months, not minutes. One line in your GitHub Action. No screenshot quotas.

$1 is a fully refundable deposit · No credit card for the free tier · Cancel anytime

[ Demo GIF placeholder ]
Diff review UI — grouped diffs on a real PR.
Replace with product mock (label it "mock", that's fine).

Screenshots evaporate.
Regressions creep back at 2am.

AI writes UI 10× faster — and ships visual bugs 10× faster too.

🧠

Eyes without memory

Your agent can verify what it just built — but what about the next PR? The next branch? The staging deploy at 2am? Screenshots evaporate; regressions creep back in. A one-off check is not a system of record.

🧾

Quotas punish the workflow AI created

Percy's free tier is 5,000 screenshots a month. Ship 10 PRs a day across 5 pages × 3 viewports and you're done by week two. AI assistants made you iterate 10× faster — per-screenshot billing taxes exactly that.

🔊

"Free" costs you in noise

Playwright screenshots plus a loose threshold catch big layout shifts — and drown you in font-rendering and animation false positives. No review UI, no PR comments, no baseline management across branches and OSes.

Eyes are free now. Memory isn't.

Agent skills answer "is this change OK?" — once. UICanary answers "did every release stay correct?" — always.

UICanaryPercyAgent skills
(ProofShot / pursr)
Playwright DIY
Answers"Did every release stay correct?"Same"Is this change OK?" (one-off)"Here's a PNG, you figure it out"
Baselines across branchesYesYesNoDIY
Multi-viewport in one runYesYesManualDIY
Team review workflowYesYesNoNo
Noise management over timeYesPartialNoDIY
Screenshot quotasNone — pay as you goYesNoneNone
Starting price$19/mo (early bird)Free 5k/mo, then quote-basedFreeFree
Honest guide: use agent skills when you want a quick check right after the agent writes code — they're free and good at that. Use UICanary when visual correctness has to hold across every PR, branch and release. Prices checked Oct 2026; competitors' plans change — verify before you buy.

Live in 5 minutes

One line in your workflow. Baselines from the first run.

1

Add one line to your workflow

Paste the Action, add your API key as a secret. Done in 5 minutes.

2

We render every PR across your viewports

Desktop, tablet, mobile — baselines stored per branch, history kept.

3

Get a PR comment — and a history you can audit

Diffs grouped by what actually changed: review 6 real changes instead of 60 screenshots. See when something broke, not just that it broke.

# .github/workflows/visual.yml
- name: Visual regression
  uses: uicanary/visual-regression@v1  # placeholder action name
  with:
    api-key: ${{ secrets.UICANARY_API_KEY }}
    viewports: "1280x800, 390x844"
    threshold: 0.1

Built to remember

Everything a one-off screenshot check can't do.

🧠 A memory for your UI

Baselines stored per branch, with full run history — audit when anything broke, not just that it broke.

🗂️ Grouped PR diffs

Built for AI-iteration workflows — diffs clustered by what actually changed, so you review 6 real changes instead of 60 screenshots.

🔇 Noise-resistant thresholds

Per-project tuning for fonts, animations, timestamps and dynamic content — false positives go down over time instead of piling up.

📐 Multi-viewport & breakpoints

Desktop, tablet, mobile in one run.

♾️ No quotas, pay as you go

Scale with your PR velocity, not against it.

🔐 Your data, your retention

Encrypted in transit and at rest; configurable retention; export or delete baselines anytime. Removing us is one line in your workflow file.

Pricing that scales with you

Start free. Lock in early-bird pricing for life before launch.

Free

$0/mo
  • 1,000 screenshots / mo
  • 1 project
  • 7-day baseline retention
  • Community support
Start free

Free tier opens at launch — join the waitlist.

EARLY BIRD — $1 PRESALE

Pro

$19/mo, locked for life
  • Unlimited projects
  • Pay-as-you-go screenshots, no quotas
  • Grouped PR diffs + full baseline history
  • Multi-viewport & breakpoints
  • Priority support
Reserve with $1

$1 is a fully refundable deposit. If we don't launch, you get it back — no questions asked.

Pro

$38/mo
  • Everything in Early-bird Pro
  • At the regular price, after launch
Join the waitlist

Launches after the early-bird window closes.

FAQ

Straight answers, no marketing fog.

They're great at point-in-time verification: "did the agent's last change look right?" UICanary is what runs on every PR in CI, keeps baselines across branches, and gives your team one place to review. Think of it as: the agent skill is eyes, UICanary is memory. If a quick check after each agent run is honestly enough for you, stick with the free tools.
Playwright gives you PNGs. UICanary gives you the other 80%: a review UI, PR comments, cross-OS baseline management, and noise thresholds — the parts that take weeks to build yourself and die to maintain.
One rendered viewport of one URL in one CI run.
Screenshots are encrypted in transit and at rest, with configurable retention and one-click delete. We never train models on your UI.
Set per-project thresholds for fonts, animations and timestamps; approve a baseline once and only real changes flag after that.
Your baselines export as PNGs anytime, and removing the Action is one line in your workflow file. No lock-in by design.

Eyes are free now. Memory isn't.

$1 reserves $19/mo for life. Refundable if we don't launch.

Lock in $19/mo for life — $1 presale

Or join the waitlist for launch updates: