What it can do:
| Feature | Example |
|---|---|
| Sentiment Analysis | Is this review positive or negative? |
| Entity Extraction | Find names, places, brands, events in text |
| Language Detection | What language is this text? |
| Topic Grouping | Automatically group articles by topic |
| Parts of Speech | Tokenize and analyze text structure |
Serverless NLP service. Finds insights and relationships inside text using ML.
Use cases: Analyze customer emails to find pain points, group news articles by topic automatically.
Comprehend Medical Same idea but for healthcare. Extracts useful info from unstructured clinical text — doctor notes, discharge summaries, test results. Detects PHI (Protected Health Information) via DetectPHI API.
Simple ML flow:
Fully managed service for developers and data scientists to build, train, and deploy ML models — all in one place.
Without SageMaker: You set up servers, install libraries, manage training jobs, deploy endpoints yourself — painful and slow.
With SageMaker: One managed platform handles everything end to end.
Use when: You need a fully custom ML model trained on your own data.
ML-powered document search service. Search across documents and get direct answers — not just a list of files.
Supported sources: S3, RDS, Google Drive, SharePoint, OneDrive, Salesforce, ServiceNow, custom sources.
Incremental Learning: Kendra learns from user interactions and feedback to improve results over time.
Use case: Internal knowledge base, HR FAQs, IT helpdesk, enterprise document search.
Automatically extracts text, handwriting, and structured data from scanned documents using AI — no manual template setup.
Works on PDFs, images, forms, tables.
Use cases:
Textract vs Rekognition:
Rekognition = analyze what is IN an image (faces, objects, scenes)
Textract = extract TEXT and DATA from a document image