AWS Textract

Analyze Document

Extract text, forms (key-value pairs), tables, and signatures from a document. Supports JPEG, PNG, PDF, and TIFF. For multi-page PDFs, use Start Document Analysis instead.

Detect Document Text

Extract plain text from a document. Faster and cheaper than Analyze Document — use this when you only need the text content without forms or tables. Supports JPEG, PNG, PDF (single page), and TIFF.

Analyze Expense

Extract structured data from receipts and invoices: vendor name, totals, tax, line items, and more. Works best with clearly formatted receipts and invoices.

Analyze ID Document

Extract structured data from identity documents such as driver's licenses and passports. Returns fields like name, date of birth, ID number, and expiry date.

Start Document Analysis

Start an asynchronous analysis job for a multi-page document stored in S3. Returns a job ID that you can pass to the "Get Document Analysis Results" action to retrieve the extracted data once the job completes.

Get Document Analysis Results

Retrieve the results of an asynchronous document analysis job started with "Start Document Analysis". If the job is still running, the status will be IN_PROGRESS. For very large documents, results may be paginated — use the returned nextToken to fetch the next page.