Privacy
Short version: we do not keep your files, we do not train on them, and no human reviews them.
- Automatic deletion. Uploads and extracted results are deleted automatically within 1 hour of upload, by a scheduled cleanup job. Rendered page images used for OCR are deleted right after extraction.
- Excluded from backups. The upload folder is not part of any backup set — nothing we back up contains your files.
- On-server processing only. Extraction and OCR run on the same server that serves this site. No cloud OCR service, no third-party file processing.
- No training, no human review. Your files are never used to train models and nobody opens them.
- Accounts. Signing in is optional for the free tier, and required to buy pages. An account stores only your email address — plus your Google subject ID if you sign in with Google, so we can recognise you on your next visit. Your files are never tied to your account: uploads are keyed by an anonymous browser quota and deleted within 1 hour either way.
- Last sign-in. When you sign in, we record the IP address and time of that sign-in, for account security. It is never shared with anyone.
- Session cookie. When you sign in we set a signed session cookie so the wallet stays logged in. It contains no file information and expires when you sign out (or after 30 days).
- Update emails. If you join the email list, we store only the email address you gave us, to send product updates. Paid packs are already live — the list is not a pre-launch waitlist. Nothing else.
- Analytics. We use self-hosted Umami analytics (no cookies, no cross-site tracking) to count page views.
If you have questions about privacy or refunds, contact us at haiyuwrld@gmail.com.