Skip to content
Back to ProUI

Privacy

What ProUI collects, and what it refuses to

ProUI works by reading your product’s screens, so the honest version of this page has to be specific. Below is what leaves your browser, what is deliberately left behind, who else processes it, how long it is kept, and how to have it deleted.

Last updated 4 October 2026.

Who and what this covers

This statement covers proui.com (this website), the ProUI web application at app.proui.com, and the ProUI browser extension for Chrome. ProUI is the data controller for the information described here, and can be reached at privacy@proui.com. The registered legal entity behind ProUI will be named on this page before signup opens.

ProUI is not open for signup yet. There is no billing system, so no payment card, billing address or tax information is collected anywhere in the product; when paid plans launch, this page will be updated before they do.

This website

This site is a set of static files. It sets no cookies, runs no analytics, no advertising and no third-party tracking scripts, and embeds nothing from a social network. The one typeface it uses is compiled into the site at build time rather than fetched from a font service, so visiting a page here does not announce you to anyone else.

Our hosting provider processes ordinary request metadata (IP address, timestamp, the page requested, the browser’s user-agent string) in order to serve the page and to absorb attacks. We do not combine it with anything else, and we do not use it to profile visitors.

Your account

Accounts are handled by Supabase. Signing up stores your email address and a password; the password is set and verified by Supabase and is never seen by, or stored on, ProUI’s own servers. Against your account record ProUI keeps your email address, a display name and avatar image URL if those are set, and the dates the record was created and last changed.

What a capture contains

When you capture a screen, two things are stored: the screenshot, and a structural description of the page called the Page Capture Model. Alongside them ProUI keeps the page’s URL and title, the viewport size and device pixel ratio, the time of capture, the image’s dimensions, format, byte size and checksum, and which project you filed it under.

The Page Capture Model is a description of structure, not of content: element tags and ARIA roles, accessible names, headings and visible label text in bounded lengths, geometry, and a palette of the computed styles in use. It is what lets ProUI reason about a screen as a layout rather than as a picture.

What the capture deliberately leaves behind

The extractor is written to not read the things that are most likely to be someone’s personal data, and this is a property of the code rather than a promise about how it is used:

Secrets are stripped before anything is sent

Every string taken off a page is passed through a redaction filter inside your browser, before it is uploaded, and again before it is stored. What it matches is replaced with a conspicuous «redacted» marker rather than something that reads as real, and the removal is irreversible: the original text is not kept anywhere to be recovered from. The filter looks for JSON web tokens, bearer tokens, AWS and OpenAI and GitHub and Slack credentials, publishable and secret API keys, anything labelled as a password, secret, token, API key, SSN, CVV or PIN, payment card numbers (checked against the Luhn algorithm, so an order number of the same length survives), email addresses, and long opaque strings such as session identifiers and signatures.

A pattern filter is a safety net, not a guarantee. It cannot recognise a customer’s name, a medical detail or a salary, because those look like ordinary words.

A screenshot is pixels, and shows what was on the screen

The exclusions above protect the structural model. They cannot protect the image: a screenshot records whatever was visible, including real names, addresses, order details and anything else on the page. This is the single most important thing to understand about using ProUI.

So the extension ships with a redaction tool. You mark an element or draw a box over it, and the mask is applied to the live page before the screenshot is taken, which means the original pixels are never photographed at all. Where that cannot hold, the mask is burned into the image bytes as solid opaque black, not a blur or a pixelation, both of which can be undone. A mark stores the selector or the rectangle, never the text it covers.

Our recommendation is plainer still: wherever you can, capture screens that hold test data rather than a real customer’s record.

What the browser extension can see

The extension installs with permission to act on the active tab only, and that permission is minted by your click on its toolbar icon. It runs no always-on content script: nothing is injected into a page until you ask for a capture, and nothing is read from pages you do not capture. Broad site access is an optional permission, requested only if you turn on captures that need it, and you can withdraw it in Chrome at any time.

Captures waiting in the extension’s tray, and your signed-in session, are held in the browser’s own extension storage on your machine until you upload or discard them.

What the AI is sent

When you run a review, ProUI sends OpenAI a downscaled copy of each selected screenshot, the structural description derived from it, the review role and focus you chose, and any instructions you typed. A review is capped at a fixed number of screens per run.

Two things about how that request is made:

ProUI does not train any model of its own on your screens, and does not use one customer’s captures to answer another’s review.

What ProUI stores about your work

Beyond the captures themselves, ProUI keeps the things it works out from them and the things you make: per-screen understandings, review findings and the regions of a screen each finding cites, the design language measured across a product and the token decisions you record against it, your canvases, annotations and wireframes, your projects and any combinations of them, and, for each AI run, its status, model, timings, token counts and any error. That last group is how cost is accounted for and how a failed run can be explained.

Who else processes it

The services ProUI relies on, what each one processes, and where it is located.
ServiceWhat it processesWhere
SupabaseAccounts, sign-in and the application databaseUnited States (us-west-2)
Cloudflare R2Screenshot and page-model storageUnited States
RailwayHosting for the ProUI APIUnited States
Cloudflare PagesHosting for this websiteGlobal edge network
OpenAIThe AI that reviews your screensUnited States
ResendTransactional emailUnited States
Google WorkspaceEmail you send to our proui.com addressesGlobal

Each is a processor acting on our instructions, and each is bound by its own agreement with us. Because the services that store your captures run in the United States, using ProUI means they are stored and processed there. We do not sell personal information, and we do not share it for advertising.

How it is protected

How long it is kept, and how to delete it

Your captures, projects and reviews are kept until you delete them. Deleting a screen removes it from the product immediately and marks its files for destruction; a scheduled job then erases the stored bytes (the original screenshot, the downscaled copy the AI reads, and every version of its page model) after a short grace period of about a week, which exists so an accidental deletion can be undone. Deleting a project or a canvas cascades through everything filed under it.

To delete your account and everything attached to it, write to privacy@proui.com from the address on the account. Deleting the account deletes the records that belong to it (captures, projects, canvases, review runs and all), and the stored files follow on the same grace period. Deleting a project or a canvas on its own is narrower: the review runs made from it survive, detached, because a run is also the record of what it cost.

Your rights

Depending on where you live, you may have the right to ask what we hold about you, to have it corrected or deleted, to receive a copy in a portable form, to object to or restrict a particular use, and to complain to your data protection authority. Reviews can be exported as PDFs from inside the product at any time. For anything else, write to the address above and we will answer within the time the applicable law allows. We will not charge you for asking, and we will not treat you differently for having asked.

Children

ProUI is a tool for product teams and is not directed at children. We do not knowingly collect information from anyone under 16; if we learn that we have, we will delete it.

Changes to this statement

This page is part of the product’s source, and it is revised when the product’s behaviour changes, not on a schedule. The date at the top is the last revision. If a change materially widens what we collect or what we do with it, we will say so directly rather than quietly updating the date, and the previous version remains in the public history of the site’s repository.

Contact

Questions, requests and complaints all go to the same place: privacy@proui.com.