AI builds the first draft. Monet runs the next hundred changes.
Capture any website on your iPad. Mark it up with Apple Pencil. Hand your agent a work order it can execute, not a screenshot it has to guess at. Review cycles get shorter because the agent spends its time changing the site instead of interpreting your feedback.
Built for iPad · desktop coming soon
Built for Claude · Codex · Cursor · and custom AI agents | Patent pending
The iteration problem
AI makes the first version instant. The next hundred changes are where projects stall.
The first draft is easy
AI ships a working site in minutes. Getting from draft to done is the hard part.
The next hundred edits are hard
Real projects need dozens of small corrections: copy, layout, visuals, brand. Each one matters.
Context gets lost
Feedback scatters across screenshots, Slack, email, and half-remembered prompts. None of it reaches your agent.
AI agents need structured handoffs
An agent can only act on what you give it. It works best with constraints, context, and clear acceptance criteria.
Monet keeps the human review loop organized.
The workflow
Ship designs in cycles.
Five steps from raw AI output to a closed loop. The compounding value shows up around cycle three, when prior rounds have cleared the easy fixes and you're working on the hard problems.
Capture
Point Monet at any deployed site. It walks the routes it finds, takes full-page screenshots, and indexes the DOM. Add any paths it misses. Sign-in walls included.
Annotate
Draw boxes with Apple Pencil. Write what you want changed and how you'll know it's done. Tag by category: copy, layout, accessibility, SEO, structure.
Hand off
One tap builds a structured packet: screenshots, crops, selectors, computed styles, acceptance criteria. Preflight catches gaps before you submit, so nothing arrives half-specified.
Reply
Your agent reads the packet and ships changes. One paste closes every matching annotation. No screenshots back and forth, no re-explaining.
Verify
Monet re-captures the site and diffs every annotated region, so a fix only counts as done when the pixels actually changed. Anything unchanged goes back to the agent.
"Capture. Annotate. Hand off. Reply. Verify."
Inside the app
The review layer your AI agent has been missing.
A purpose-built iPad workspace: site map, annotated captures, and structured handoff, all on one screen.
What Monet doesn't do
It doesn't write code.
Your agent makes the changes in your repo. Monet writes the instructions and checks the results, and that separation is deliberate: the tool that grades the work should not be the one doing it.
It doesn't phone home.
Captures and annotations stay on your device. There is no account, no cloud sync, and no ad or tracking SDKs riding along. Client work stays yours.
It doesn't take "fixed" on faith.
Every agent says fixed. Monet re-captures the page and diffs the exact region you circled. That closed loop is the patent-pending part.
Pricing
Two apps, one loop.
Start on iPad: capture, annotate, and hand off. The desktop app closes the loop with reply and verify. The apps are sold separately, each through its own store and with its own free trial.
Monet for iPad
First to ship7-day free trial. Cancel anytime. Built for iPad, on iPadOS 26 and later.
- ✓Apple Pencil and finger annotation, with notes and acceptance criteria
- ✓Unlimited projects and captures, stored on device
- ✓Structured handoff bundles for Claude Code, Codex, Cursor
- ✓DOM-anchored crops, selectors, and computed styles per note
- ✓Sign-in capture for gated and client sites
- ✓No account, no cloud, no data collection
Pair with Monet for Desktop to close the loop with reply and verify.
Coming soonjoin the launch list
Monet for Desktop
Coming soon14-day free trial. Cancel anytime. Native app for macOS and Windows 11.
- ✓The full solo loop: capture, annotate, hand off, reply, verify
- ✓Verify-the-fix re-capture with per-region diffs
- ✓Pointer and keyboard driven, on macOS or Windows 11
- ✓Same portable bundles, same local-first privacy
- ✓Signed installer with automatic updates
Sold separately from the iPad app, through Paddle.
Get notified at launchiPad subscription terms. Subscriptions renew automatically unless cancelled at least 24 hours before the period ends. Payment is charged to your Apple ID at confirmation. Any unused portion of the free trial is forfeited when you subscribe. Manage or cancel anytime in your Apple ID settings. By subscribing you agree to our Terms and Privacy Policy.
Desktop subscription terms. Desktop purchases will be processed by Paddle, our merchant of record. After the 14-day free trial, the subscription renews automatically at the price shown until you cancel. Cancel anytime from the link in your receipt email or by writing to support@iammonet.com. The iPad and desktop apps are separate purchases.
Bring your own key. The optional AI Design Critic runs on your own Anthropic key, usually pennies per page. The key stays in your device Keychain. We never see it and never take a cut of your AI spend.
FAQ
Common questions.
Which coding agents does it work with? +
Claude Code, Codex, Cursor, or any agent that reads files or accepts pasted context. Monet produces a portable bundle: markdown plus structured JSON plus screenshots. No lock-in to one AI vendor.
Do I install anything on my website? +
No. Monet captures sites from the outside, the way a visitor sees them. It works on any stack: WordPress, Shopify, React, plain HTML, anything with a URL.
Where does my data live? +
On your device. Captures, annotations, and bundles stay local. No account to create, no cloud to configure. Subscriptions are handled by the App Store on iPad, and by Paddle on desktop.
Does it work on sites behind a login? +
Yes. Sign in once inside Monet's capture browser and it replays that session for future captures, including sites gated by a PIN or a client-side login.
What about Mac and Windows? +
A native desktop app for macOS and Windows 11 is in active development. It runs the full loop, including the reply and verify steps, for $9.99 a month or $99 a year after a 14-day free trial. Join the launch list and you'll hear about it first.
Do I need separate subscriptions for iPad and desktop? +
Yes, for now. The iPad app is sold through the App Store and the desktop app is sold through Paddle, so each has its own subscription and its own trial. We may link them in the future.
Who makes Monet? +
Garton Holdings, LLC, a California company. The capture-annotate-handoff-verify pipeline is patent pending (U.S. provisional 64/071,601).
The story
A name with roots. A tool with purpose.
Ben Garton
Founder, iammonet.com
linkedin.com/in/benwgarton- Agricultural automation and robotics
- AI-driven development tools
- Monet family lineage research
The name is personal.
Monet is not just a product name, it is a family name. Ben Garton's great-grandmother always told the family they were related to Claude Monet, the Impressionist painter. That story was not just oral tradition. It was documented.
In 1981, Montez DeMonia Jones published A DeMoney-DeMonia Family: Demoneia, Monet, Money, Monnet, LeMoine (Gregath Co., Cullman, Alabama, OCLC 8629280), a genealogical record tracing the descendants of Henry de Money and the related Monet, Monnet, Money, and DeMoney branches of the family tree. The book is held in WorldCat and stands as a published record of the lineage.
The name Monet, then, carries both Impressionist resonance and genuine family history, a tool named not just for an aesthetic, but for a bloodline.
A DeMoney-DeMonia Family
Demoneia, Monet, Money, Monnet, LeMoine
Author: Montez DeMonia Jones
Publisher: Gregath Co., Cullman, Alabama, 1981
OCLC 8629280, WorldCat genealogy record
Built from the field up.
Ben has spent years deeply involved in agricultural automation and robotics, systems where precision, iteration, and human oversight are not optional. In the field, a machine that acts without human confirmation is a liability. The same principle applies to AI-built software.
As AI coding tools matured and became capable of generating entire websites in minutes, Ben recognized a gap: the tools were fast, but the review process was not. Feedback was scattered, context was lost, and AI agents were receiving vague instructions that produced vague results.
Monet is the tool Ben needed. A structured, visual, human-in-the-loop review layer that turns impressions into implementation, the same discipline that makes agricultural automation reliable, applied to AI web development.
"AI builds the first draft. Monet runs the next hundred changes."
Get the launch email.
We send two emails: one when Monet reaches the App Store, and one when the desktop app ships. That is the whole list.
painted, not promised