Skip to content

Capture

Documents into your form
with the phone camera.

Your customers photograph a vehicle registration, ID or receipt, and your web form fills itself. One script tag on your website, no app, processed only in the EU.

From the form to the phone and backcapture.js
  1. 01The customer clicks the button and scans the QR code
  2. 02Photograph and check, right on the phone
  3. 03The fields in the form are filled in

How it works

Scan a document with the phone, fill the form automaticallyThree steps. No typing.

Fewer typos, fewer abandoned forms: the values come straight from the document.

1 · The button in your form

Your visitor clicks “Scan vehicle registration”. On a laptop a dialog with a QR code opens; on a phone the camera starts straight away.

2 · Photograph and check

A frame guides the photo. Blur and glare are detected before sending, so a bad photo is retaken right away instead of failing later.

3 · The form is filled in

The read values land in the fields you mapped in the portal, on the device where the form was opened. Your visitor checks them and sends the form as usual.

Use cases

Read vehicle registrations, IDs and receipts in web formsWherever forms ask for documents.

Every document type of the API works, including full text. A form can ask for several documents in a row.

Insurance and car dealers

Vehicle registration for car insurance, trade-ins or registrations: registration number, VIN, holder, first registration, HSN/TSN.

Rentals and customer onboarding

ID document for rental agreements or new customer accounts: name, date of birth, document number and validity.

Expenses and refunds

Receipts and invoices for expense claims or refund forms: supplier, date, amounts and taxes.

Applications with proof

Payslips, bank statements or contracts as proof in an application, several documents in one form.

Integration

One script tag, mapping in the portalReady in an afternoon.

Create a form in the portal: documents, appearance, which value goes into which field, and the websites it may run on. Paste the snippet. Your server can confirm every result with the session id and token from the form.

Capture documentation
<script src="https://capture.deep-ocr.com/v1/capture.js"
        data-capture-id="cap_live_…" async></script>

<form>
  <button type="button" data-deepocr-open>Scan vehicle registration</button>
  <input name="registration_number">
  <input name="vin">
</form>

Confidentiality

Document capture processed in the EUPhotos that do not linger.

Capture follows the same rules as the API: processing in the EU, clear deletion, no training. Contract documents for your data protection review are available on request.

Processing only in the EU

Photos go from the phone straight to Deep-OCR and are read in data centres within the EU.

Deleted after reading

Photos are stored encrypted only until they are read, then deleted. Unfinished sessions expire after 30 minutes.

No model training

Neither photos nor read values are used to train or improve AI models.

Only on your websites

A live id works only on the websites you approve in the portal, over https.

Pricing

Test Capture for freeFree to test. Live on request.

Test

Free with any account: create a form in the portal and try it right away, on your own computer or on the portal's test page with your phone.

Live on your website

An add-on that we enable on request. Live sessions are billed like any extraction, by pages read. Abandoned sessions cost nothing.

Good to know

FAQ about CaptureClear answers. Before you start.

Do my customers need an app?

No. Capture runs in the browser. On a laptop the dialog shows a QR code that opens the camera on the phone; on a phone the camera starts directly. The photos go straight from the phone to Deep-OCR, and the fields appear in the form on the device where it was opened.

Which documents can be captured?

Every document type of the API, for example vehicle registrations, ID documents, invoices, receipts, contracts or payslips, and also full text. A form can ask for several documents in a row, each with one or more pages, and documents can be optional.

What happens to the photos?

They are processed only in the EU and stored encrypted until they are read. A document's photos are deleted as soon as it has been handed over for reading, and the extracted values are available for a short time only. Unfinished sessions expire after 30 minutes and are deleted together with their photos. Nothing is used to train models.

How do I check the result on my server?

The form receives a session id and a token together with the values. Your server sends both to the API and gets the read result back, so values typed in by hand can be told apart from values read from the document.

What does Capture cost?

Testing is free with any account: create a form in the portal and try it on your own computer or with your phone on the portal's test page. Live use on your own website is an add-on that we enable on request. Live sessions are billed like any extraction, by pages read; abandoned sessions cost nothing.

Let's talk about your form

Capture documents in web formsFewer fields to type. More completed forms.

Tell us which documents your form asks for. We look at the fields, the flow and the data protection requirements together with you.