Amazon Comprehend

Machine Learning

Managed NLP service that pulls meaning — sentiment, entities, language, PII — out of unstructured text.

Verified against AWS documentation · 2026-08-24

Key facts

  1. 01Send text, get back structured output: sentiment, entities, key phrases, dominant language, PII — no model to train
  2. 02Two modes: real-time APIs for one document or a batch of 25, and async jobs that read a whole corpus from S3
  3. 03Custom classification and custom entity recognition train on your own labels — you supply examples, not algorithms
  4. 04Limits: ~5,000 bytes per document on most real-time calls, 25 documents per batch, topic modeling is async-only

Exam triggers

Phrases and scenarios that point here.

  • "Analyze customer feedback", "gauge sentiment", "extract insights from text", "no ML expertise on the team"
  • Detect or redact PII inside free-form text before storing or sharing it → Comprehend, not Macie
  • Scanned forms or PDFs → Textract first, Comprehend second — Comprehend only reads text, never images
  • Call recordings or voice notes → Transcribe first, Comprehend second — audio is never Comprehend's input

Not this service

  • Text trapped in scanned documents, forms, or images → Textract
  • Sensitive data sitting at rest in S3 buckets → Macie
  • Clinical notes, PHI, medication or ICD-10 extraction → Comprehend Medical

Use cases

  1. 01Support ticket triage — custom classification routes tickets to the right queue, sentiment flags the angry ones for escalation before the SLA clock runs down
  2. 02Scrubbing PII out of chat and email transcripts before they land in a data lake or get handed to an external analytics vendor
  3. 03Review mining at scale — entities plus key phrases across thousands of reviews to find which product features people actually talk about, and targeted sentiment to tell whether each one is praised or complained about
  4. 04Multilingual help desk — detect the incoming language, route to the right agent pool or hand off to Translate, all before a human sees the message
  5. 05Topic modeling over a legacy document archive nobody has read in years, to find out what's actually in there before setting a retention policy

How it's charged

  1. 01Priced per 100-character unit with a 3-unit (300 character) minimum per request — a 20-character request costs the same as a 300-character one, so batch short inputs
  2. 02Rates differ by feature family: sentiment, entities, and key phrases sit on one tier, PII detection on another
  3. 03Topic modeling is billed per job based on total input size, not per unit
  4. 04Custom models cost three ways: training time, monthly model storage, and inference
  5. 05Real-time custom endpoints bill per inference-unit-hour from provisioning until deletion, traffic or not — an idle endpoint left up over a weekend is the classic surprise line item

Commonly confused with

Amazon TextractMachine LearningAmazon MacieSecurity, Identity & ComplianceAmazon TranslateMachine Learning

References