All AI news

AWS announces Traditional Chinese print recognition for Textract; handwriting remains English-only

Automatic text detection updates retain separate language limits for Forms, Queries and other features.

AWS announced on November 2, 2026 that Textract’s DetectDocumentText and AnalyzeDocument APIs automatically gain text detection in over 40 languages, including Traditional Chinese, without request or response changes. It says the update is available wherever Textract operates; no exact activation time is given.

The official FAQ now lists over 40 languages for printed text, but Forms extraction remains limited to six languages; handwriting, Queries, invoice and identity-document analysis remain English-only. Some official sections still retain the older language list, and the limits page excludes vertical text. Usage is charged, with conditional free allowances. We have not tested it.

Official FAQ: https://aws.amazon.com/textract/faqs/
Limits: https://docs.aws.amazon.com/textract/latest/dg/limits-document.html
Pricing: https://aws.amazon.com/textract/pricing/
Endpoints: https://docs.aws.amazon.com/general/latest/gr/textract.html

Usage conditions

Hong Kong
Hong Kong: No local deployment is listed among Textract endpoints. Individual access to supported overseas regions remains unverified and does not imply local processing. This is an API update, with no separate mobile-app announcement. Traditional Chinese support concerns printed text; analysis features have separate limits. Per-page and feature charges depend on the selected region.
Macau
Macau: No local deployment is listed among Textract endpoints. Individual access to supported overseas regions remains unverified and does not imply local processing. This is an API update, with no separate mobile-app announcement. Traditional Chinese support concerns printed text; analysis features have separate limits. Per-page and feature charges depend on the selected region.
Taiwan
Taiwan: No local deployment is listed among Textract endpoints. Individual access to supported overseas regions remains unverified and does not imply local processing. This is an API update, with no separate mobile-app announcement. Traditional Chinese support concerns printed text; analysis features have separate limits. Per-page and feature charges depend on the selected region.

Official source: Announcing an updated text detection model for Amazon Textract

Source published:

Source checked:

Published here: