Skip to main content

Amazon Textract

See Amazon Textract Developer Guide for general guidance using Textract.

API Access​

The ITextractClient interface and its implementation TTextractClient provides access to all Textract operations.

program ExtractText;

uses
  AWS.Textract;

var
  Client: ITextractClient;
  Request: ITextractDetectDocumentTextRequest;
  Response: ITextractDetectDocumentTextResponse;

begin
  Client := TTextractClient.Create;

  Request := TTextractDetectDocumentTextRequest.Create;
  Request.Document := TTextractDocument.FromFile('Path\\To\\Image.jpg');
  try
    Response := Client.DetectDocumentText(Request);
    // Inspect Response which includes text Blocks and DocumentMetadata.
  except
    on E: ETextractException do
      Writeln(E.ClassName, ': ', E.Message);
  end;
end.

Analyzing Documents​

Input documents can be analyzed for relationships between detected items using AnalyzeDocument. Alternatively, documents can be analyzed asynchronously using StartDocumentAnalysis and the results can be fetched at a later time with GetDocumentAnalysis.

Analyzing Expense Documents​

Expense related information can be extracted from invoices/receipts using AnalyzeExpense. Alternatively, expense documents can be analyzed asynchronously using StartExpenseAnalysis and the results can be retrieved later using GetExpenseAnalysis.

Analyzing Lending Documents​

Lending documents can be analyzed using the StartLendingAnalysis function. Results can be retrieved later using GetLendingAnalysis.

Analyzing Identity Documents​

Identity documents can be analyzed using AnalyzeID.

Detecting Text​

Text can be detected in an input document using DetectDocumentText. Text can be extracted asynchronously using StartDocumentTextDetection and the results can be retrieved later using GetDocumentTextDetection.