Loading…

Features Pricing

Log in Get started

Bounding Box

Bounding Box

Concepts

A bounding box is a set of coordinates that define a rectangular region on a document page—often page index plus x, y, width, height (or equivalent). In DocLD, bounding boxes can be returned with extraction citations or parsing output so you can highlight "where" in the document a value or passage came from. They support visual citation and overlay in viewers.

Use Cases

Citation — Show the exact region that supported an extracted field or a RAG answer.
Overlay — Draw highlights or boxes on a PDF or image viewer.
Validation — Inspect the source region when confidence score is low.

Bounding boxes are derived from parsing and layout analysis; accuracy depends on document structure and OCR quality for scanned documents.

Related Concepts

Bounding boxes accompany citation and extraction results. They come from parsing and layout analysis. Confidence score and confidence threshold can be used with regions to flag review.

Related terms

Citation
Extraction
Parsing

See also

features › extraction

Frequently Asked Questions

A bounding box is a set of coordinates (e.g., page, x, y, width, height) that define a rectangular region on a document page. DocLD can return bounding boxes with extraction citations so you can highlight the exact source region.

Use bounding boxes to show where in the document an extracted value or RAG citation came from, to draw highlights in a PDF viewer, or to validate low-confidence extractions by inspecting the source region.

Typically normalized (0–1) or in points. Check the API response for the coordinate system so you can map to your viewer.

Bounding boxes are optional and may be present for extraction or RAG when the pipeline supports it. Page and text excerpt are more commonly always present.

Enable or request bounding box output in the extraction config if supported. Not all document types or parsers support coordinates.

Map the coordinates to your viewer's coordinate system and draw a rectangle or highlight. Different viewers use different origins and units.

Yes. OCR and layout analysis can produce coordinates for text regions. Quality depends on the scan and OCR output.

A field might have one box per occurrence or one for the whole value. Schema and API define whether multiple boxes are returned.

Fall back to page number and text excerpt. Some pipelines do not produce coordinates; your UI can still show page and quote.

See the DocLD documentation and glossary for details, and check the API reference for related endpoints and options.

On this page

Use Cases
Related Concepts
Frequently Asked Questions

Product

Features
Pricing
API Reference

Industries

Insurance
Construction
Professional Services
Finance
Energy & Utilities
View all

Company

About
Careers

Resources

Documentation
Blog
Help Center
Status

Legal

Privacy Policy
Terms of Service
Trust

Connect

X
GitHub
Discord

© 2026 DocLD, Inc.SOC audit in progress