Features

The Vidado READ API enables organizations to extend their existing document digitization capabilities and include low quality, low resolution and even handwritten information. Through the READ API, simply supply an image containing text, such as a snippet of a field from a form, and the READ API will return back the estimated value and corresponding confidence score.

Anyone can register and request access to the READ API at https://api.vidado.ai/portal, but all requests will be screened prior to providing API keys. Each key will enable a developer to submit one shred image per second to the READ API, and up to 1,000 shreds per day. Images of up to 1MB in size and up to 100,000 square pixels in size are supported. 

In this initial release, the READ API only supports text digitization. This includes alpha-numeric strings, dates, and numbers. Other field types such as signatures, checkboxes and radiobuttons are not currently supported.

The READ API is ideal for customers, partners and solutions integrators that already have some form of documentation classification and digitization capabilities, and are looking to augment their system with improved low quality, low resolution or handwritten capture.


Benefits

The READ API can be used to "read" handwritten or machine-printed text data from an image and return the digitized text.

  • Image formats: Accepted image formats are jpeg, jpg, gif, png
  • Character limit: The max number of characters extracted from the supplied image is 1024 characters
  • Casing: The response extracted text will be in lower case. The READ engine is case agnostic.
  • Vertical text: The API doesn't rotate or align the image and cannot read vertical text.
  • Text skewing: The underlying machine learning model that the API uses was trained against +/- 5 degrees of text skewing to help ensure accurate transcriptions.
  • Storage: The API doesn't store/save the supplied images.
  • Character set: The API can transcribe numeric and alphanumeric characters, it cannot transcribe accents or roman numerals.
  • Training: No training is necessary. The Machine learning model that is used to deliver the results is training is pre trained by Vidado and deployed ready for use.
  • Machine Only: Vidado READ is a pure machine only solution with no humans in the loop. 

Resources

Inputs

Pre-requisite:

Step to get digitized data:

  • An image snippet corresponding to the area of a field on a form.
  • Make a REST API call to submit image for digitization.
  • Receive digitized data in response.

Outputs

  • Digitized text value
  • Confidence score generated by Vidado READ for the digitized text. 

Please refer to the sample documentation to get greater technical insight. 


Geographic Availability

Global

Language Availability

English