Hyperclay Privacy Policy
Effective Date: March 30, 2026
What We Collect
- Account info: email, username, password (hashed).
- Payment info: processed by Stripe. We store only your Stripe customer and subscription IDs, never card numbers.
- Usage data: server logs, plus the visit statistics described in Visit Statistics below.
Visit Statistics
- Why we count visits: We count visits to our public websites (hyperclay.com's marketing pages, htmlclay.com, clayjs.com, and over time our other sites such as hyperclaylocal.com and saphtml.com) using PostHog, whose servers for our project are in the USA. The purpose is UX improvements only: understanding where people get confused or stuck, so we can make the products easier to use.
- What we count on public pages: We count pages viewed, how far down a page people scroll, clicks (including combined click and scroll heatmaps), clicks that do nothing, errors, page loading speed, and a few named actions such as clicking a download link or copying a code snippet. Query strings in page addresses (for example tokens in links) are removed before anything is sent, except campaign tags such as utm_source.
- What we never do: We never use session recordings, never record or replay what any individual visitor sees or does, and never sell this data. Analytics never loads inside anyone's documents or published sites, and never in the ClayJS script that documents load.
- Identifiers and IP addresses: Visitors get a random identifier stored in their browser (cookie and local storage), and we never link it to a name, email, or account. IP addresses are used only to look up the country and are then discarded, not stored.
- Who is not counted: We do not count visitors in the European Union, the European Economic Area, the United Kingdom, or Switzerland at all, and nothing is sent to PostHog from those visitors. Browsers that send Global Privacy Control are never counted.
- When a visit counts: A visit is only counted after the visitor actually scrolls, clicks, taps, or types. Automated browsers and known crawlers are not counted.
- Logged-in dashboard: In the logged-in dashboard, for paid accounts only, we record a few named actions (created a document, opened a document in the editor, uploaded a file, invited a teammate, created a share link, started an AI build, saw an upgrade prompt, and saw an error as an error code), and dashboard addresses are reduced to a single generic path. We never send document contents, file names, page titles, or anything typed.
- Feedback answers: If someone answers the optional one-question feedback prompt from our mascot on the pricing and download pages, their answer is stored in PostHog with the visit.
What's Public
- A public document can be read by anyone (see Terms, Section 1).
- A private document opens for you, your team, the people you share it with, and anyone holding a share link you created. Hyperclay administrators can also open it when running the service.
- Uploaded files, such as images, can be opened by anyone who has their address, even when the document that shows them is private. Don't upload secrets.
- When you upload an HTML file without an account, its preview link works for one hour only in the browser that uploaded it. We keep the file privately for 30 days and then delete it unless you sign up to keep it in your account.
How We Use Your Data
- To operate your account and process payments.
- To improve the service.
- We don't sell your data. We share it only with the companies below, to run the service, and when required by law.
Companies That Handle Your Data
- Hetzner runs our server in Ashburn, Virginia, USA.
- Amazon S3 stores our nightly backups.
- Stripe processes payments.
- Postmark sends the email we send you.
- Google Workspace receives the email you send us.
- PostHog counts visits on our public websites, records named actions in the paid dashboard, and stores optional feedback answers on servers in the USA.
- Cloudflare serves the ClayJS scripts that make documents editable. When anyone imports a page, a Cloudflare worker reads its text, headings and form fields once so Workers AI can suggest editing tools, and nothing is stored. For free accounts only, uploaded pages also pass through a Cloudflare worker to stop phishing and abuse.
- DeepInfra receives a page's address, file name and HTML when you run Add CMS on it. It receives your instruction and the HTML of the element you are editing when you use AI edit. On top of that, it receives any text you selected inside that element. If you reference the whole page with @page, and only then, it also receives the page's full HTML up to 400 KB. When you use page tools, it receives your request, a short outline with the page's title, its headings and the shape of its tables, and the tool's current code if you revise one. When you use Describe anything with generated follow-up questions on, it receives your description.
- Anthropic receives your description and your answers to the guided or generated follow-up questions when you use Describe anything. It also receives any suggested answers you kept and any short comments you added.
Where Data Is Stored
- Our server is in the USA. See Security for how documents, versions and backups are kept.
Deleting Your Account
- Email privacy@hyperclay.com to request deletion.
- You have 30 days to undo by logging back in.
- You choose: keep your published sites live (read-only) or take everything down.
- After 30 days, we permanently erase your email, password, and payment info.
- We keep a one-way hash of your email so you can restore your account if you sign up again later with the same address.
- Usernames remain reserved after deletion.
Your Rights
- You can export your whole workspace as a zip at any time from the account menu.
- You can request a copy of your data or ask questions at privacy@hyperclay.com.