Amazon Textract extracts text and structured data from scanned documents. Beyond basic optical character recognition (OCR), it can identify relationships in forms and the structure of tables, helping applications process invoices, forms, identity documents, and other paper-based records.
What makes Textract different?
- Text detection: identify lines and words in a document image.
- Forms: extract key-value relationships such as a field label and its value.
- Tables: detect table structure and cell relationships.
- Specialized workflows: use supported analysis features for common document types such as invoices and receipts.
Textract vs. Amazon Augmented AI
Textract performs the document analysis. Amazon Augmented AI (A2I) helps add human review to machine-learning predictions when a workflow needs people to validate low-confidence or sensitive results. These services can complement each other: Textract extracts data, and a human-review workflow can help verify results where required.
Exam decision clues
Choose Textract when the requirement is to extract text, form fields, or table data from scanned documents. Do not select Transcribe, which handles audio speech, or Translate, which translates text between languages.
Related topics
Review Amazon Augmented AI (A2I) and the AWS Topics study guides.
Prepare for the AWS Certified AI Practitioner (AIF-C01)
Review AWS AI services and document-processing use cases with original practice questions.
Try the free diagnostic or explore the 390-question premium practice package.