
Our OCR assistance features enhance contributor efficiency by up to 4.75x.
- OCR pre-labeling predicts bounding boxes and transcriptions for every word in a document.
- Predict and transcribe in 34 languages including right-to-left languages.
- Real-time OCR provides contributors a prediction immediately upon placing a bounding box. They can then edit the predicted transcription.

Capture the relationships needed to make your document data actionable with our intuitive relationship annotation feature.
- Map relationships across PDF pages.
- Flexible and configurable.
- Annotate key-value pairs.

With our flexible solution, create the relationships needed to capture tables.
- Associate columns and rows with parent tables.
- Associate column or row cells with parent headers.
- Export parent-child relationships in JSON format.

The varied shapes of handwriting, logos and expressive fonts require more than four vertices to localize for the greatest accuracy.
- N-sided polygon annotation is supported in addition to bounding box annotation.
- Specify the maximum number of vertices allowed per polygon.

Use our tool’s transcription validators to prevent transcription errors from occurring for fields with known character rules, for the greatest accuracy.
- Validate dates, alphabetical characters, alphanumeric characters, decimals, and integers.
- Validate character counts.
- Allow or disallow symbols.
- For fully custom validations, specify the rules with regex.