Skip to content
OCR-AI
Open a document

Security and privacy

What happens to a document, step by step

The document opens in your browser, one page at a time is sent to be read, and the data comes back to you. Along the way nothing is kept: not the file, not the image, not the text. This page sets out exactly what travels, what is stored and where.

  1. 01

    Your browser

    The file opens here. A PDF itself never leaves your computer: only an image of the page being read.

  2. 02

    The OCR-AI server

    Checks your page allowance and passes the page on to be read. It does not keep the image, the text or the file name.

  3. 03

    The reading service

    A cloud service on a paid plan whose terms forbid using the content to train models.

Page imageThe data

01What is stored

What is stored, and where

With us, on the server

  • An anonymous browser ID, in a signed cookie, to count free pages.
  • Daily page counters: per browser, per network (a keyed hash of the IP address, not the address itself) and for the whole site.
  • For each page read: when, the file size, how long it took, the document type (invoice, receipt…) and whether it succeeded.
  • For people who bought pages: the email from checkout, the page balance and the purchase record at Stripe.

Never stored with us

  • The file you uploaded.
  • Page images.
  • The text and the extracted data.
  • The file name.
  • Card details.

Only in your browser

  • Your document history: page images and results, in this browser's IndexedDB. Up to the 30 most recent documents.
  • Nobody else sees it, us included. Delete it with one click, below or in the workspace.

02The reading service

Who reads the page

The reading itself is done by an external cloud AI service. The page is sent to it as an image, and it returns text, fields and tables.

We use it on a paid plan whose terms forbid using submitted content to train models. Under its terms, it may keep a request for a limited time to monitor for abuse, and for no other purpose.

On our server the page exists in memory only, for the duration of the request, and is gone when the request ends. System logs record usage without content: times, sizes and error codes.

03Payment and account

Payment, sign-in and cookies

  • Payment at Stripe

    Purchases happen on Stripe's checkout page. Card details never pass through us and are not stored by us.

  • No passwords

    You sign in with a one-time link sent to your email. We store only a hash of the link; it is valid for 20 minutes and for one use.

  • Signed cookies

    The browser and sign-in cookies are cryptographically signed, readable only by the server (HttpOnly), and sent only over secure connections.

  • Rate limits

    Reading and sign-in requests are rate-limited, and once a day you may be asked to confirm you are not a bot (Cloudflare Turnstile).

04Limits

Worth knowing before you upload

  1. 01

    Pages are sent to an external cloud service to be read. Do not upload documents you are not allowed to share with a cloud provider.

  2. 02

    Results are produced automatically and can be wrong. Values the engine was unsure of are marked, but check every value before using it in your books.

  3. 03

    History is stored in the browser. Anyone using the same computer and browser profile can see it. On a shared computer, delete it when you are done.

  4. 04

    The site runs measurement tools (Google Analytics, Cloudflare and a visit-measurement system) that see which pages were viewed. They do not receive document content or file names. Details are in the cookie policy.

05Control

Delete, ask, report

Want your account deleted (the email and the page balance)? Write to us and we will delete it.

Found a security problem? We want to know. Write to us and we will get back to you.

Contact form