Inside DocVault
Built into DocVault document management. It can mask files automatically in chosen folders, at upload, as a workflow step, or as a bulk job over existing files.
30 capabilities, each labelled so you know what works today, what's available on request, and what's still planned.
Where it runs, and how it reaches your documents.
Built into DocVault document management. It can mask files automatically in chosen folders, at upload, as a workflow step, or as a bulk job over existing files.
Run the engine as a container in your data centre or cloud, so images never leave your network. It runs on ordinary CPUs; no GPU is needed.
Inside DocVault, bulk jobs mask whole folders of files you already hold, including every stored version, and report each skipped file with the reason.
Walk through the API flow your systems would use, on your own card, with no sign-up: a key, an upload, the status and the masked result, with access rules applied.
What happens to an image, and what it hides.
The first eight digits of the Aadhaar number are covered with a solid block. The last four stay readable, in the format of a masked Aadhaar.
The QR code on the card carries personal details too, so it is found and covered completely.
A trained vision model finds the Aadhaar card in the image, including a card that is part of a larger photo or scan.
Two text-recognition engines read the card and pick out the 12-digit number, printed horizontally or vertically.
Sideways, upside-down and tilted photos are straightened before they are read.
A card that already shows a masked number is recognised and not masked a second time.
An image with no Aadhaar number comes back byte-for-byte unchanged, and you're told so.
A JPEG comes back as a JPEG and a PNG as a PNG, with the rest of the image untouched.
What you can send.
Phone photos and scans, checked by their real content, not their name.
The engine also handles TIFF images and PDF documents, page by page, when you run it yourself or inside DocVault.
Empty, truncated and damaged files, and files pretending to be images, are refused before processing.
What happens to your data.
Your image is held in memory only, never saved to disk or a database.
Each masked result is released once, to the request that made it, then erased. Uncollected results are erased automatically after the time you set.
We never store the Aadhaar number (not even the last four), your file name, or anything read from the card.
Every request is logged with its time, source, file type and size, and outcome, never the image or anything read from it.
How applications and people use it.
Send an image, check its status, and collect the masked copy once, with a clear result: masked, no Aadhaar found, or a reason it was refused.
Set the rules for your deployment: key expiry, images per key, requests per client in a time window, file size and resolution caps, and how long uncollected results are kept. Keys are tied to the client that requested them, and only a one-way hash of each key is stored.
A bounded queue in front of the engine. Callers see their place in line, and a full queue answers with how long to wait.
Runs on ordinary CPU servers. Tested on CPU-only hardware: an Intel Core i7 machine (4 cores, 8 threads, 32 GB of memory) and a 2-vCPU, 8 GB cloud server.
Not available yet. Tell us which ones matter to you.
Multi-page PDFs and TIFFs through the online service, not only when self-hosted.
Your own API credentials, higher limits, usage reporting and an SLA for your apps.
Send many images in one request and collect the results together.
A callback to your system when a result is ready, instead of checking its status.
An extra check on every masked image, which flags any image where a full number might still be readable.
Faster masking on GPU servers for high volumes.
Exploring masking for other Indian ID documents.
No feature matches that search. Try a broader word, or ask us.
Book a demo on your own documents, or ask our sales team about pricing.