What ScanCortex does with your documents

Where your data goes

What we store

Nothing. There are no accounts, so there is no email address, no username and no profile. Your documents, the pages inside them and the text read out of them are written to ScanCortex's own storage area on your phone and nowhere else. They are not backed up to us, because there is no "us" to back them up to.

ScanCortex contains no analytics, no crash reporting and no tracking library of any kind. Apart from Google's advertising SDK, described below, the only request the app ever makes is the one that sends a page to be read, and it makes that only when you ask it to.

The app also keeps a few small preferences on the device: your display language, your light or dark theme, which reading language you last picked for Extract text, whether you have seen the welcome screen and a hint in Settings, and a counter the app uses to space ads out. None of them identifies you, and none of them leaves the phone.

Text recognition

Saving a scan or importing a file does not read it and sends nothing: the document is stored on the phone with no text. A page is read only when you ask — Extract text on a document, or PDF to Word in Tools.

Extract text reads on our server unless you tell it otherwise. Its reading-language picker starts on Detect the language, which sends the page to our server, works out the script there and reads it. That is what covers the long list of languages in the picker, and while it is selected every page you ask to read is sent to us. For a PDF, its pages are turned into images on the phone and sent one at a time, up to the first thirty. PDF to Word always reads on our server.

For Extract text you can choose to read on the phone instead. The picker also offers five on-device scripts — Latin, Chinese, Japanese, Korean and Devanagari — and picking one of those runs Google ML Kit entirely offline: the page never leaves the device and no network request is made. There is one exception. If an on-device read comes back empty, or comes back looking like the wrong script, the app sends that page to the server to try again, so that a page in a script your phone cannot read still gets read.

When a page is sent, the image is held in memory only for as long as the reading takes — it is never written to disk, never stored in a bucket, and no copy is kept. The server either reads it itself, or passes it to OCR.space or Google Cloud Vision, which return the text and are bound by their own policies. Only the image of the page being read is sent — never the document file itself, and never the text already read from other pages.

Two small things travel with that page. The first is its file name, which for a scan is something like page-1.png, but for a file you imported is the name you gave it. The second is the language the app itself is set to, sent as a hint about what the page is likely to say. Both are discarded with the page.

Because that endpoint is open and each reading costs money, our server counts requests per IP address for a rolling sixty seconds and refuses more than twenty. The count is held in memory and expires with the window; it is not written down, not linked to anything, and not used for any other purpose.

Advertising

We show ads through Google AdMob. AdMob may use your device's advertising identifier to select ads and measure how they perform. It does not receive your documents, your scans, or the text in them — none of that is ever sent to an advertiser.

On iPhone, on first launch iOS will ask whether ScanCortex may track you across other companies' apps and websites. If you decline, you will still see ads; they will simply be less relevant. You can change your answer at any time in Settings › Privacy & Security › Tracking. On Android, the advertising identifier is controlled in your phone's settings (see Turning off personalised ads below). Where the law requires it, we also ask for consent before showing personalised ads.

Permissions

The camera is used only while you are actively scanning. Photo library access is used only when you pick an image to import, and only for the images you choose. You can also import a file from Files, which hands us only the file you picked. Nothing is captured in the background, and none of these is used for anything else.

Deleting your data

Deleting a document removes it from your phone. Settings › Storage › Delete All Documents does the same for every document at once. Deleting the app removes everything it ever stored.

Because nothing of yours is held anywhere else, that is genuinely all there is to delete — there is no server-side account to close and no request you need to send us.

Your rights

Because ScanCortex has no accounts, there is very little to exercise a right against. The only things of yours that ever reach us are the pages you have text read from, each held in memory and then discarded, and the IP address they arrived from, held for under a minute. We cannot look you up, because there is nothing to look you up by.

In the EEA and the UK

You have the right to access your personal data, correct it, delete it, restrict or object to how it is processed, and receive it in a portable form. Our legal bases are narrow: reading a page you asked us to read is performance of the service you requested; counting requests per IP address is our legitimate interest in stopping an open, paid-for service from being abused; and personalised advertising happens only with your consent, which we ask for when you first open the app. You may lodge a complaint with your local data protection authority at any time.

In California

We do not sell your personal information for money. Allowing tracking does let Google use your device's advertising identifier for cross-context behavioural advertising, which California law treats as sharing — declining the tracking prompt on iPhone, or deleting your advertising ID on Android, opts you out of it. We do not knowingly sell or share the personal information of anyone under 16.

Turning off personalised ads

On iPhone, open Settings › Privacy & Security › Tracking and turn ScanCortex off. On Android, open Settings › Google › Ads (on some phones Settings › Privacy › Ads) and tap Delete advertising ID. Ads keep working; they simply stop being tailored to you.

Who else is involved

• Google ML Kit On-device text recognition, when you choose it. Runs offline; sends nothing.
• Google AdMob Advertising, and the only recipient of an advertising identifier.
• OCR.space and Google Cloud Vision Read the pages sent to our server, one page at a time.
• Vercel Hosts the reading service. Like any web host it processes the connection itself, including the IP address it arrives from.

Children

ScanCortex is not directed at children under 13, and we do not knowingly collect their data. Since the app collects no personal information from anyone, there is nothing collected from a child either.

Changes to this policy

If we change how any of this works, we will update this page and the copy inside the app, and change the effective date at the top. Significant changes will be called out in the app itself.

Contact