Link Search Menu Expand Document

Product FAQs

  1. Document Solutions
    1. Extraction
    2. Masking
    3. Verification with External Databases
  2. Face Solutions
    1. Face Compare
    2. Face Validation (Liveness)
  3. RFT (Right First Time) SDK
  4. Miscellaneous

Document Solutions


What is extraction?

Extraction is a type of task, where our APIs read the text on ID card images & tag them. We use OCR technology to read the text & ML based classifiers to tag the data.

What are the common use-cases for the OCR service?

The common use-cases are -

  • Assisting manual data entry by pre-filling the form fields
  • Ensuring that the document provided & the data entered by the user are matching

What is the average response time does an OCR request take?

The processing time for an OCR request is generally less than 5 seconds. The processing time may increase in the following scenarios due to extra processing required -

  • Image is not horizontally aligned. We first compute the rotation angle, rotate the image, and then start the OCR
  • Image is not in JPG/PNG format. We convert the images in other formats to JPG/PNG for performing the OCR
  • Image is of poor quality. We do some image processing to ensure better OCR results

I already have my user’s data. I want to match the document with the available data. Is there a way to do that?

Yes, you can do that. In the request body, provide the available data along with the document image. We will provide a match result in the response. Match results are only provided for name, father’s name & address. For more details, please have a look at the API spec.

Do I need to specify the document type for OCR?

Yes, you need to select your task type based on the document. For example, there are separate task types for DL OCR & Voter Id OCR. If you select the task type as Voter Id OCR, and provide a DL Image, the request will be marked as invalid.

How do you validate a document while doing OCR?

In OCR, the input image is validated using text markers that are present on the ID cards. The basic validation in OCR is used to discard wrong ID images or images where the text is not readable, however it is not completely foolproof.

Do you provide a confidence score with your OCR output?

No, currently we do not provide a confidence score in the OCR output

Can I provide both front & back images of the document in a single OCR task?

Yes, you can either provide either of the following in a single OCR task -

  • Both the front and the back images
  • Only the front image
  • Only the back image

You do NOT need to specify the side of the image (front or back), our machines can detect that. Please note that you will be charged for every task, so we recommend that you send both the front & back images in a single task.


What is masking?

Masking is a type of task to mask some sensitive data on a document image. We currently offer masking only for Aadhaar card, where we mask the first 8 digits of the aadhaar number.

  • The input is an image with the complete aadhaar card
  • The output is the link to an image with masked aadhaar number
  • The masked image is provided in less than 8 seconds, and the solution works on ID card images rotated at any angles

Verification with External Databases

Which documents can be verified using databases? Which databases are used?

We can currently verify 3 types of Indian ID documents. The checks are automated. -

  • Individual PAN with NSDL database
  • Company PAN with NSDL & MCA database
  • Driving Licence with Saarthi database
  • Voter Id with Electoral Database (NVSP)

Face Solutions

Face Compare

What is Face Compare API?

Face compare is a task type that compares 2 face images, and provides the match result. We currently provide 4 types of responses based on the thresholds that we have set -

  • Faces match, no manual review required
  • Faces match, manual review recommended
  • Faces do not match, manual review recommended
  • Faces do not match, no manual review required

Can the Face Compare APIs be used for comparing a selfie with image on ID card?

Yes, this is the primary use-case for our face compare APIs (for authentication). Our APIs work well in the following scenarios -

  • Document face image is clear but of poor quality
  • There is a difference in faces due to aging, facial hair, etc

Besides this, our face compare APIs are also used for following use -

  • Doc1 face Image vs Doc2 face image
  • Selfie1 vs Selfie2

Face Validation (Liveness)

What is Face Validation API?

Face validation is a task type that checks if a live face is present in an image. Our APIs detect liveness from a single image. The primary use-case is to detect identity spoofing attempts where someone uses faces opened on a screen or faces printed on a paper.

RFT (Right First Time) SDK

What is the “Right First Time” SDK?

RFT SDK is an android SDK to capture images of documents & faces right first time. It has functionalities to capture & upload the captured images.

  • Capture right documents - Ensure that documents are captured correctly by identifying unreadable & incorrect images immediately. The errors in the capture are immediately highlighted to the users via a review screen
  • Capture right selfie - Ensure that the face image is captured correctly by face detection, face orientation detection, closed eyes detection, multiple-face detection and some other features. The errors in the capture are immediately highlighted to the users via a review screen
  • Upload images - Upload the images directly to IDfy servers & get an image link. This image link can then be used to create Tasks from the server


Do you provide on-premise solutions?

No, we currently only provide cloud based solutions