What's the difference between cloud-based and on-device data extraction?

Cloud-based and on-device data extraction represent two different architectural approaches to processing document images, and the right choice depends on a business's specific requirements around connectivity, privacy, and processing power.

On-device extraction runs the entire OCR, MRZ reading, and validation process locally on the user's phone or the business's scanning hardware, without sending the document image to an external server. This offers a few concrete advantages.

It works without an internet connection, which matters in settings like remote areas or certain airport environments where connectivity can be unreliable. It's generally faster, since there's no network round-trip involved in getting a result.

And it offers stronger privacy guarantees by default, since the sensitive document image never needs to leave the device unless the business specifically chooses to transmit the final extracted data elsewhere afterward.

Cloud-based extraction sends the captured image to a remote server, which performs the OCR and validation processing and returns structured results back to the requesting app or system. This approach can draw on more computationally intensive processing than a phone's own hardware might support, and it centralizes model updates and improvements, meaning the OCR engine can be upgraded on the server side without requiring every user to update an app on their device.

The tradeoff is a dependency on network connectivity for every scan, and a different privacy posture, since the document image is transmitted to and processed by a server outside the user's own device, even if that transmission and any temporary storage are properly encrypted.

Many vendors offer both options, or a hybrid model where some processing (like initial image capture guidance) happens on-device while more intensive analysis happens server-side, giving businesses a choice depending on their specific connectivity, privacy, and processing requirements rather than being locked into a single model.

ScanDoc supports both approaches, letting businesses choose on-device processing for scenarios prioritizing offline capability and stronger built-in privacy, or cloud-based processing where more centralized, server-side capability better fits their operational setup.