OCR (Text Recognition)

Overview

Hive's text recognition models detect and transcribe each word in an image. They also return semantically grouped and ordered text blocks (Block Text) in their natural reading order for detections that are spatially close. For details on our OCR moderation API, please see the OCR (Text Recognition) Moderation tab.

For documents that have a structure, such as receipts, Hive provides APIs that interpret the structure — such as returning each item on a receipt, its unit price, quantity, and total price.

Note: these models are optimized for images with 150 words or fewer. If you submit images with more words than that, we recommend you split the image into multiple segments and submit them separately.

Block Text: "Enterprise Al solutions to power the next wave of automation"Block Text: "Enterprise Al solutions to power the next wave of automation"

Block Text: "Enterprise Al solutions to power the next wave of automation"

Language Support

Our models currently support the following languages:

  • English
  • Spanish
  • Portuguese
  • German
  • French
  • Italian
  • Chinese (simplified + traditional)
  • Thai
  • Russian
  • Japanese
  • Korean
  • Indonesian
  • Malay
  • Swahili
  • Hindi
  • Arabic
  • Greek (coming soon)

If you are unsure if your required language is supported/want to request an additional language, please reach out to our sales team ([email protected]).

Relevant Domains

We've trained our models across a wide variety of domains, including but not limited to the following categories:

  • Rotated Text
  • Handwritten Text
  • Memes
  • Receipts
  • Menus
  • Scene Text (text in photographs)
Handwritten TextHandwritten Text

Handwritten Text

Scene TextScene Text

Scene Text

Rotated TextRotated Text

Rotated Text