The Comprehend connector allows processes to analyze text using Amazon Comprehend natural language processing (NLP) services.
Use the Comprehend connector to extract entities, identify personally identifiable information (PII), classify documents, and perform text analysis on documents and text content. The connector supports standard entity recognition, custom entity recognition, and custom document classification models.
The connector is displayed as a rectangle with the Comprehend logo:
The connector submits text or document content to Amazon Comprehend for analysis. Amazon Comprehend processes the content and returns analysis results, which may include extracted entities, PII matches, classification results, or confidence scores. The returned information can be mapped to process variables and used by gateways, forms, integrations, and other process elements.
The connector can analyze plain text provided directly to the connector or files containing text content.
The connector supports a wide range of text-based document formats, including:
- TXT
- HTML
- XML
- CSV
- Microsoft Word documents
- Microsoft Excel workbooks
- Microsoft PowerPoint presentations
- OpenDocument files
- Apple Pages, Numbers, and Keynote files
- ZIP and archive formats
To configure the Comprehend connector:
-
Configure AWS. For more information, see Configuring AWS for the Comprehend Connector.
Note: The connector requires access to an AWS account with permission to invoke Comprehend functions. - Place the connector in your process diagram.
-
Fill the connector's configuration parameters to create the authentication. For more information, see Comprehend Connector Configuration Parameters.
The Comprehend connector authenticates using AWS credentials configured through connector configuration parameters. The configured AWS user or role must have permission to access Amazon Comprehend services and any custom models used by the process.
- Configure the connector action.
The connector supports the following actions:
| Comprehend Connector Action | Description |
|---|---|
| ENTITY | Identify entities such as people, organizations, locations, and dates within text. |
| PII_DETECTION | Detect personally identifiable information in text content. |
| DOCUMENT_CLASSIFICATION | Categorize documents using custom Amazon Comprehend classifiers. |