# Amazon Comprehend

Cloud service · AWS. Updated 2026-11-12. Page: https://blowstack.com/tools/cloud-services-hub/amazon-comprehend

> Amazon Comprehend is a natural language processing (NLP) service that uses machine learning to uncover insights and relationships in text, enabling developers to easily integrate data analysis and language interpretation capabilities into their applications.

Practice exams covering Amazon Comprehend: [AIB-C01 (beta) AWS Certified AI Business Strategist](https://blowstack.com/practice-exams/aws-certified-ai-business-strategist), [AIF-C01 AWS Certified AI Practitioner](https://blowstack.com/practice-exams/aws-certified-ai-practitioner), [CLF-C02 AWS Certified Cloud Practitioner](https://blowstack.com/practice-exams/aws-certified-cloud-practitioner), [AIP-C01 AWS Certified Generative AI Developer - Professional](https://blowstack.com/practice-exams/aws-certified-generative-ai-developer-professional), [MLA-C02 (beta) AWS Certified Machine Learning Engineer - Associate](https://blowstack.com/practice-exams/aws-certified-machine-learning-engineer-associate), [SAA-C03 AWS Certified Solutions Architect - Associate](https://blowstack.com/practice-exams/aws-certified-solutions-architect-associate).

## 11 key features of Amazon Comprehend

Amazon Comprehend is a comprehensive Natural Language Processing (NLP) service that offers entity recognition, sentiment analysis, language detection, key phrases extraction, text extraction from custom documents, and the ability to create custom classification models.

1. **Natural Language Processing (NLP)**: Amazon Comprehend uses advanced machine learning models to analyze text and understand its meaning, making it easier to extract insights, relationships, and sentiment.
2. **Entity Recognition**: Automatically identifies entities such as people, places, brands, or products, and labels them for easy analysis, helping enhance content organization and retrieval.
3. **Sentiment Analysis**: Determines the sentiment expressed in a body of text, classifying it as positive, negative, neutral, or mixed, which can be pivotal for customer feedback and social media monitoring.
4. **Language Detection**: Quickly identifies the language of the text from a broad range of languages, supporting multi-language applications and content analysis without language pre-specification.
5. **Key Phrases Extraction**: Extracts key phrases, or the most relevant terms from a piece of text, making it easier to understand the main points and summarizing content efficiently.
6. **Text in Custom Documents**: Enables the extraction and analysis of text from a variety of document formats, allowing for the processing of different types of written content, such as PDFs or Word documents.
7. **Custom Classifications**: Allows for the creation of custom text classification models tailored to specific needs, facilitating the categorization of text according to predefined labels or categories.
8. **Targeted Sentiment**: Determines sentiment for specific entities mentioned in a document, with each mention classified as positive, neutral, negative, or mixed.
9. **Event Detection**: Detects specific types of events and their related details in documents.
10. **Flywheels**: Flywheels help manage custom classification and custom entity recognition models.
11. **AWS KMS Encryption**: Works with AWS Key Management Service (AWS KMS) to provide enhanced encryption for your data.

## Where Amazon Comprehend is used

Amazon Comprehend is used for sentiment analysis, entity recognition, language detection, key phrases extraction, and text classification, aiding businesses in understanding, organizing, and acting on large volumes of text data.

1. **Sentiment Analysis**: Amazon Comprehend can be utilized to analyze text documents or files stored in Amazon S3 to determine the overall sentiment expressed within them. This is particularly useful for understanding customer feedback on products, services, or brand sentiment on social media, allowing businesses to quickly identify and focus on positive or negative feedback for further action.
2. **Entity Recognition**: This functionality enables the extraction of entities such as persons, places, brands, or products from text. Businesses can leverage this for enhancing content personalization, improving search and discovery mechanisms, or for automating data entry processes by extracting relevant information from unstructured data.
3. **Language Detection**: Amazon Comprehend's language detection feature is capable of accurately identifying the language of the text documents. This is crucial for global businesses that operate across multiple regions, allowing them to route content to the appropriate language-speaking teams for further analysis, translation, or direct engagement.
4. **Key Phrases Extraction**: This use case involves extracting key phrases from text, making it easier to understand the main points or topics being discussed. Businesses can use this feature for content summarization, enhancing metadata for content categorization, or optimizing search engine marketing strategies by identifying relevant keywords.
5. **Text Classification**: Amazon Comprehend can classify text into predefined categories or classes, facilitating the automation of document sorting, email routing, or content moderation. This allows businesses to streamline operations by ensuring that the content is quickly and efficiently categorized without manual intervention.

## How Amazon Comprehend is priced

Amazon Comprehend's pricing models include an On-Demand, pay-as-you-go model based on text volume and feature usage, alongside a specific, higher-priced model for medical text analysis.

- **On-Demand Pricing**: Amazon Comprehend uses a pay-as-you-go pricing model based on the amount of text processed and features used. Prices vary for different functionalities, such as entity recognition, sentiment analysis, etc., and are measured in units of 100 characters of processed text.
- **Comprehend Medical**: Separate pricing applies for Comprehend Medical, Amazon's dedicated service for medical text analysis. Costs are higher compared to standard features and are calculated based on the amount of text processed, aiming to provide in-depth medical insights.
- **Custom Model Training**: Training custom classification and entity recognition models costs $3 per hour, billed by the second.
- **Custom Model Management**: Custom model management costs $0.50 per month.
- **Provisioned Custom Inference**: Synchronous custom classification and entity recognition endpoints cost $0.0005 per inference unit per second. Each inference unit provides 100 characters per second of throughput. Billing uses one-second increments with a 60-second minimum and continues until the endpoint is deleted, even when no documents are analyzed.
- **Topic Modeling**: Topic modeling is charged by the total document size processed per job, with a flat charge for the first 100 MB and a per-MB charge above 100 MB.
- **Free Tier**: The free tier provides 50,000 units of text per eligible API per month and five topic modeling jobs of up to 1 MB each. It is available to new and existing AWS customers for 12 months from their first Amazon Comprehend request. Custom model training, inference, and management are excluded.

## Services Amazon Comprehend works with

- **[Amazon S3](https://blowstack.com/tools/cloud-services-hub/amazon-s3)**: Amazon Comprehend can read documents stored in Amazon S3 for natural language processing.
- **[Amazon CloudWatch](https://blowstack.com/tools/cloud-services-hub/amazon-cloudwatch)**: Amazon Comprehend integrates with Amazon CloudWatch for monitoring and logging of events and metrics.
- **[AWS Lambda](https://blowstack.com/tools/cloud-services-hub/aws-lambda)**: Amazon Comprehend can be used with AWS Lambda to create automated workflows, such as triggering data analysis upon file upload.
- **[Amazon Connect](https://blowstack.com/tools/cloud-services-hub/amazon-connect)**: Amazon Comprehend provides natural language processing for Amazon Connect conversational analytics, helping identify sentiment in customer interactions.
- **[Amazon EventBridge](https://blowstack.com/tools/cloud-services-hub/amazon-eventbridge)**: Amazon Comprehend API calls recorded by AWS CloudTrail can trigger Amazon EventBridge rules. For example, a rule can route a custom-model creation request to a notification or processing workflow.
- **[Amazon Lex](https://blowstack.com/tools/cloud-services-hub/amazon-lex)**: Amazon Comprehend powers sentiment analysis in Amazon Lex, allowing a bot to receive sentiment scores for user utterances. These scores can help a conversation flow respond differently when a user expresses frustration.
- **[Amazon QuickSight](https://blowstack.com/tools/cloud-services-hub/amazon-quicksight)**: Amazon Comprehend is available through action integrations in Amazon Quick Suite, the expanded offering that includes QuickSight. Its actions can classify individual documents using a custom model endpoint and extract key phrases from text.
- **[Amazon SageMaker Ground Truth](https://blowstack.com/tools/cloud-services-hub/amazon-sagemaker-ground-truth)**: Amazon Comprehend integrates with Amazon SageMaker Ground Truth to label training documents for custom entity recognition. Annotators can mark domain-specific entities, such as product identifiers, to produce training annotations for a recognizer.
- **[Amazon Textract](https://blowstack.com/tools/cloud-services-hub/amazon-textract)**: Amazon Comprehend uses Amazon Textract to extract text and document structure from supported image and PDF inputs for custom classification and entity recognition. Document-reader settings control how extraction is performed, including whether to extract forms and tables.
- **[Amazon Translate](https://blowstack.com/tools/cloud-services-hub/amazon-translate)**: Amazon Translate uses Amazon Comprehend to detect the source language when a translation request specifies automatic language detection. This lets callers translate text without supplying its source language.
- **[Amazon VPC](https://blowstack.com/tools/cloud-services-hub/amazon-vpc)**: Amazon Comprehend supports VPC configurations for custom-model training and analysis jobs, using network interfaces in specified subnets and security groups to access resources privately. Its APIs can also be reached through interface VPC endpoints.
- **[AWS Amplify](https://blowstack.com/tools/cloud-services-hub/aws-amplify)**: Amazon Comprehend powers text interpretation in AWS Amplify Predictions. Applications can use this integration to detect sentiment, entities, key phrases, language, and syntax in supplied text.
- **[AWS CloudFormation](https://blowstack.com/tools/cloud-services-hub/aws-cloudformation)**: Amazon Comprehend custom document classifiers, entity recognizers, and flywheels can be provisioned as AWS CloudFormation resources. Templates define their configuration so they can be deployed and managed alongside related infrastructure.
- **[AWS CloudTrail](https://blowstack.com/tools/cloud-services-hub/aws-cloudtrail)**: Amazon Comprehend integrates with AWS CloudTrail to record API activity, including the caller, request time, and request parameters. These records support auditing of model management and document-analysis requests.
- **[AWS Identity and Access Management (IAM)](https://blowstack.com/tools/cloud-services-hub/aws-identity-and-access-management-(iam))**: Amazon Comprehend uses IAM policies to control access to analysis APIs, custom models, endpoints, and other resources. Jobs also use IAM service roles to access input documents, training data, and output locations in Amazon S3.
- **[AWS Key Management Service (AWS KMS)](https://blowstack.com/tools/cloud-services-hub/aws-key-management-service-(aws-kms))**: Amazon Comprehend integrates with AWS KMS to encrypt supported job storage volumes, model data, and analysis outputs using customer-managed keys. Access policies must allow Comprehend to use the keys required by the job.
- **[AWS Organizations](https://blowstack.com/tools/cloud-services-hub/aws-organizations)**: Amazon Comprehend supports AWS Organizations AI services opt-out policies. Organization administrators can use these policies to prevent eligible Comprehend content from being used to improve AWS AI services across selected accounts.
- **[AWS PrivateLink](https://blowstack.com/tools/cloud-services-hub/aws-privatelink)**: Amazon Comprehend supports AWS PrivateLink interface endpoints, providing private API connectivity from a VPC without an internet gateway or NAT device. Endpoint policies can further restrict which Comprehend requests are allowed through the endpoint.
- **[AWS Step Functions](https://blowstack.com/tools/cloud-services-hub/aws-step-functions)**: Amazon Comprehend APIs can be invoked through AWS Step Functions AWS SDK integrations. A state machine can start an asynchronous analysis job, check its status, and continue processing after the job completes.

## What is Amazon Comprehend

Amazon Comprehend is a natural language processing (NLP) service that uses machine learning to uncover insights and relationships in text. Launched by Amazon Web Services (AWS), this powerful tool is designed to read, comprehend, and analyze vast amounts of textual data, enabling developers and businesses to understand and use the information contained within documents, social media, webpages, and more, without the need for deep machine learning expertise.

One of the key features of Amazon Comprehend is its ability to perform sentiment analysis. This enables users to determine the sentiment expressed in a body of text, whether it's positive, negative, neutral, or mixed, making it an invaluable asset for businesses looking to gauge public opinion about products, services, or brand image.

Additionally, Comprehend can be used for entity recognition, which involves identifying key elements such as locations, people, brands, or dates within a text, offering the potential to automate various data extraction tasks and enhance data analysis processes. Moreover, Amazon Comprehend supports language detection, capable of identifying the dominant language from a plethora of textual sources. This feature is particularly useful in today's global digital landscape, where data often spans multiple languages.

The service also offers syntax analysis, dissecting texts to understand the roles played by words and their relationships to one another, providing deeper insights into the grammatical structure of sentences. Another significant capability of Amazon Comprehend is key phrase extraction, which identifies the main points in a body of text. This can dramatically improve content analysis and information retrieval, helping users quickly understand the essence of documents without reading them in their entirety.

Additionally, the service includes the ability to detect personally identifiable information (PII), helping organizations protect privacy and comply with data protection regulations by identifying sensitive information within datasets. Amazon Comprehend is not just powerful but also highly accessible and easy to integrate into existing applications, thanks to its API-based interface. Users do not require machine learning expertise to leverage Comprehend's capabilities, as it offers pre-trained models that can instantly be applied to text. Furthermore, it can process thousands of documents in a multitude of languages, making it scalable to the needs of any project, from analyzing a few documents to processing millions of pages.

For synchronous inference with custom classification and custom entity recognition models, you provision an endpoint with the throughput your application needs.

Amazon Comprehend supports synchronous processing for one document or a batch of up to 25 documents, and asynchronous jobs for large document sets.

Each asynchronous API operation has a requests-per-second quota per Region and account, and supports a maximum of 10 active jobs.

In conclusion, Amazon Comprehend stands out as a sophisticated, yet user-friendly service that democratizes the power of natural language processing. By providing detailed insights and understanding from unstructured text data, it enables enhanced decision-making, streamlined processes, and a deeper understanding of content for businesses and developers across diverse industries. Whether it's through sentiment analysis, entity recognition, or any of its other features, Amazon Comprehend is crucial for anyone looking to harness the potential of NLP in the modern data-driven landscape.

## Concepts related to Amazon Comprehend

- [Data Anonymization](https://blowstack.com/tools/cloud-concepts-hub/data-anonymization): Data anonymization transforms information so that individuals are no longer reasonably identifiable while preserving useful characteristics for analysis and AI.
- [Data Labeling](https://blowstack.com/tools/cloud-concepts-hub/data-labeling): Data labeling is the process of assigning meaningful annotations or target values to data so that machine learning systems can learn from or be evaluated against known examples.
- [Data Minimization](https://blowstack.com/tools/cloud-concepts-hub/data-minimization): Data minimization is the practice of collecting, processing, sharing, and retaining only the data necessary for a defined purpose.
- [Deduplication](https://blowstack.com/tools/cloud-concepts-hub/deduplication): Deduplication identifies repeated representations of the same data and removes or consolidates them according to explicit matching and retention rules.
- [Intent Classification](https://blowstack.com/tools/cloud-concepts-hub/intent-classification): Intent classification identifies the goal expressed in a user's message and maps it to a defined category that guides the application's next action.
- [Machine learning](https://blowstack.com/tools/cloud-concepts-hub/machine-learning): Machine learning is a branch of artificial intelligence that uses data to train models to make predictions, recognize patterns, or generate outputs without explicitly programming every decision rule.
- [Metadata Extraction](https://blowstack.com/tools/cloud-concepts-hub/metadata-extraction): Metadata extraction is the process of identifying and capturing structured attributes from documents, their contents, and their source systems.
- [Named Entity Recognition](https://blowstack.com/tools/cloud-concepts-hub/named-entity-recognition): Named Entity Recognition identifies spans of text that refer to entities, such as people, organizations, locations, and dates, and assigns them predefined categories.
- [Natural Language Processing](https://blowstack.com/tools/cloud-concepts-hub/natural-language-processing): Natural language processing is the field of artificial intelligence that enables computers to analyze, interpret, and generate human language.
- [Privacy-Enhancing Technologies](https://blowstack.com/tools/cloud-concepts-hub/privacy-enhancing-technologies): Privacy-enhancing technologies are techniques that reduce the exposure of sensitive information while allowing data to remain useful for analysis, collaboration, and AI.
- [Supervised Learning](https://blowstack.com/tools/cloud-concepts-hub/supervised-learning): Supervised learning is a machine learning approach that learns from labeled examples to predict outcomes for new inputs.
- [Text Normalization](https://blowstack.com/tools/cloud-concepts-hub/text-normalization): Text normalization converts equivalent textual representations into consistent, task-appropriate forms while preserving the meaning needed by downstream systems.
- [Unstructured Data](https://blowstack.com/tools/cloud-concepts-hub/unstructured-data): Unstructured data is information whose meaningful content does not fit a predefined tabular schema, such as free-form text, images, audio, and video.