Scan data
We process data needed to run scans, build reports, prevent abuse, and debug failures.
- Submitted URL, normalized target, scan profile, report language, and report visibility.
- Limited HTTP evidence: status codes, selected headers, discovered public files, relevant URLs, and short snippets.
- Temporary operational metadata: timestamps, request IDs, errors, rate-limit and security events, including request IP address when needed to protect the service.
Privacy boundaries
The service is designed for public pages. Do not submit URLs with credentials, secrets, personal tokens, or private parameters.
- The MVP scanner does not execute remote JavaScript, log in to websites, or run user journeys.
- Full HTML is not stored by default. Evidence is reduced to data needed for checks and result explanation.
Account data
If you create an account we store what is needed to sign you in, keep the session alive, and investigate a sign-in you did not make. Everything below is optional: it exists only for people who chose to register.
- Your email address, your interface language, and the display name you chose, if any.
- Your password is stored only as an argon2id hash. We never hold it in a readable form, which is also why we cannot send it back to you and can only offer a reset link.
- One record per signed-in device: the browser identification string it sent, when the session started and was last used, and a keyed hash of the IP address it came from. The address itself is not stored, and the hash cannot be reversed without a key held separately from the database.
- A security trail of account events: registration, verification, successful and failed sign-ins, sign-outs, password changes, session revocations, and lockouts. It records hashes and event types, never addresses or raw IPs.
- Deleting your account overwrites its personal data, ends every session immediately, and frees the email address to register again. Scans you started are kept as anonymous scans, so that reports you shared do not break, and they stop naming you.
Personalized AI Fix Plans
If you explicitly request a personalized Fix Plan, Anthropic processes a bounded projection of the deterministic report to rewrite its remediation guidance for your selected role, goal, technical level, and editable surfaces.
- Anthropic receives structured findings, closed reason codes, and guidance authored by aiSiteReady. Personalization answers are closed choices with no free-text field.
- We exclude the scanned URL and query, raw page content and HTML, cookies, credentials, access tokens, secrets, raw headers, and private infrastructure values. aiSiteReady does not persist prompts or generated plans.
- On Anthropic's side, the report projection we send and the generated response are retained for 30 days and may be accessed within that window for safety and security purposes.
- For generated plans, a privacy-safe generation ledger keeps model and usage metadata plus the current structured feedback verdict for 30 days. No free-text feedback is collected or stored.
Browser and service storage
Reports and evidence are stored to display results, export reports, support repeat viewing, debug failures, and enforce abuse controls.
- The browser may store the selected interface language and private report access tokens in localStorage; the temporary admin token and a closed-choice AI personalization form draft are stored in sessionStorage.
- When you have signed in, the browser stores one flag in localStorage recording that a session may exist, so the page does not ask the server about a session you never had. It holds no token and identifies nobody.
- Public reports may be available to anyone with the link. For sensitive targets, use private visibility and do not submit secret URLs.
Cookies and analytics
We use Google Analytics 4 (GA4), loaded through Google Tag Manager (GTM), to understand aggregate, anonymous usage and improve the product. Analytics run only after you accept: Google Consent Mode v2 keeps analytics storage denied by default, and our cookie banner switches it to granted only with your consent. Analytics events are limited to coarse, non-identifying signals — page views and aggregate funnel events — and never include the scanned URL or any personal identifier.
- cc_cookie — remembers your cookie choices so the banner does not reappear. Strictly necessary; kept about 6 months (182 days).
- asr_refresh — keeps you signed in. Set only when you sign in, HttpOnly, so no script on the page can read it, and sent only to the account endpoints. Kept up to 30 days, or until you sign out. Strictly necessary; no analytics purpose.
- _ga — Google Analytics; distinguishes visitors. Set only after you accept analytics; kept up to 2 years.
- _ga_* (per-property, e.g. _ga_XXXXXXXXXX) — Google Analytics; preserves session state. Set only after you accept analytics; kept up to 2 years.
- Google Tag Manager itself sets no cookies; it only loads the consent-gated GA4 tags that set the cookies above.
Your consent choices
On your first visit, a banner lets you accept or reject analytics; strictly necessary cookies always run. You can change or withdraw your choice anytime via "Cookie settings" in the footer, which re-opens the preferences panel. If you reject or later withdraw consent, analytics storage stays denied and the _ga cookies are not set; you can also clear cookies in your browser. Google processes analytics data under its own policy (policies.google.com/privacy), and you can install its opt-out add-on (tools.google.com/dlpage/gaoptout).
Retention and deletion
MVP retention periods and deletion procedures are product operations settings and may change. Do not submit data that requires a separate contractual or regulated processing regime.