Skip to main content
Version: 8.9 (unreleased)

Amazon Textract connector

The Amazon Textract connector enables you to integrate your BPMN service with Amazon Textract to automatically extract text and data from documents.

Prerequisites​

You need an AWS IAM Access Key and Secret Key with the following permissions:

  • AmazonTextractFullAccess — required
  • AmazonS3ReadOnlyAccess — required if using S3 as the document source
  • AmazonS3FullAccess — optional if using S3 as the output location for asynchronous execution

For setup instructions, refer to the AWS Textract Developer Guide.

Authentication​

Select an authentication type from the Authentication dropdown:

  1. Credentials:
    Use if you have an AWS Access Key and Secret Key. Supported in both SaaS and Self-Managed deployments.
    Required fields:

    • Access Key: AWS access key for the user with Textract permissions.
    • Secret Key: Corresponding AWS secret key.

  2. Default Credentials Chain (hybrid/Self-Managed only):
    Use if your system relies on implicit authentication (e.g., IAM roles, environment variables, or credentials files).
    This option uses the Default Credential Provider Chain to resolve credentials.

note

Use Camunda secrets to avoid exposing token credentials as plain text. See managing secrets for details.

Configure AWS region​

  • Region: Enter the AWS region where the Textract service and your S3 buckets are hosted (e.g., us-east-1, eu-west-1).
note

Choose the region that matches your Textract and S3 resources to minimize latency and meet compliance requirements. See AWS Regions.

Configure input document​

Set the document location in the Document location type field.

S3​

  • Document Bucket: Name of the S3 bucket containing the document. Ensure proper permissions for Textract access.
  • Document path: Full path from the bucket root to the document.
  • Document version (optional): Specify if you need to process a specific document version. If not set, the latest version is used.

Camunda Document​

  • Document: Select the document from the Camunda document store.
note

Only PNG and JPEG formats are supported. Real-time execution only.

Configure operation​

Execution types​

Select an option from Execution Type:

  • Real-time:
    For small files requiring immediate text extraction. Only single-page PDFs are supported when using S3. For multi-page PDFs, use Polling or Asynchronous.

  • Polling:
    Starts analysis and polls every five seconds until the result is available. Best for larger documents where blocking execution is acceptable.

  • Asynchronous:
    For large or complex documents processed in the background.
    Configure:

    • Output S3 Bucket
    • Output S3 Prefix

    Optional fields:

    • Client Request Token
    • Job Tag
    • KMS Key ID
    • Notification Channel Role ARN
    • Notification Channel SNS Topic ARN

Feature types​

Select one or more Feature Types:

  • Analyze Tables — extracts tabular data
  • Analyze Forms — extracts key-value pairs
  • Analyze Signatures — detects signatures
  • Analyze Layout — extracts layout elements (lines, words, spatial data)

At least one feature type must be selected. Combining multiple options can produce richer extraction results.

Response​

The connector response mirrors the AWS Textract API, depending on the execution type:

note

From version 8.7.0, the Amazon Textract connector can read input directly from the Camunda document store. See document handling for details and limitations.

Use the Textract connector response in your process​

You can map fields from the Textract response to process variables using Result Variable or Result Expression.

  1. Result Variable — stores the entire response in one variable, e.g., myResultVariable.
  2. Result Expression — extracts specific fields for more granular control.

Example Textract response (real-time execution)​

{
  "DocumentMetadata": {
    "Pages": 1
  },
  "Blocks": [
    {
      "BlockType": "LINE",
      "Confidence": 99.1,
      "Text": "Hello World",
      "Geometry": {
        "BoundingBox": {
          "Width": 0.5,
          "Height": 0.1,
          "Left": 0.1,
          "Top": 0.1
        }
      },
      "Id": "abcd1234"
    }
  ]
}

Mapping example​

= {lineText: blocks[0].text, textConfidence: blocks[0].confidence, blockType: blocks[0].blockType}

Result:

{
  "lineText": "V3",
  "textConfidence": 91.60225677490234,
  "blockType": "LINE"
}