Amazon Kendra is a managed intelligent search service that answers natural language questions from an organization’s documents.
Intelligent search
Amazon Kendra searches an organization’s documents and answers questions in natural language, ranking results by meaning rather than by matching keywords.
How it works
- Create an index and connect data sources, such as Amazon S3 or Microsoft SharePoint, so Kendra crawls and ingests the documents.
- Users ask a question through the search API from a website or app.
- Kendra returns the most relevant phrase, passage or document, filtered by what that user may access.
Exam cues
- If the question says natural language search or “find answers in our documents”, think Amazon Kendra, because it is the managed intelligent document search service.
- If the question asks for a new search or RAG build after July 2026, think Amazon Bedrock Managed Knowledge Base, because Kendra closed to new customers.
Remember
- Semantic search over connected repositories; answers factoid and descriptive questions.
- Maintenance mode since June 30, 2026; closed to new customers since July 30, 2026.
Trap: Kendra finds answers in your documents. Amazon Textract extracts text from scanned documents, and Amazon Comprehend analyzes text you give it.
From the AWS docs (6)
- Amazon Kendra is a managed intelligent search and retrieval service that uses natural language processing and deep learning to rank documents and passages by semantic and contextual relevance instead of keyword matches.
- A Kendra index ingests documents from connected data sources such as Amazon S3 and Microsoft SharePoint, so one search covers several repositories.
- Kendra answers factoid questions with a word or phrase, descriptive questions with a passage or document, and conversational keyword queries by inferring their meaning.
- Kendra search results can be filtered by user or group access to documents, so people see only what they are allowed to; authenticating users stays the customer's job.
- A Kendra GenAI index serves retrieval augmented generation as well as enterprise search, and can back Amazon Q Business and Amazon Bedrock knowledge bases.
- Amazon Kendra entered maintenance mode on June 30, 2026, with no new features, and closed to new customers on July 30, 2026; AWS recommends Amazon Bedrock Managed Knowledge Base for new search applications.
Flashcards
Nothing to review here right now. Pick other topics or another filter.
·
All 4 cards as a list
Employees need to ask "Where is the IT support desk?" and get an answer drawn from internal documents in S3 and SharePoint. Which AWS service was built for this?
Amazon Kendra, a managed intelligent search service over company documents. It closed to new customers on July 30, 2026, so new builds use Amazon Bedrock Managed Knowledge Base.
How does Amazon Kendra search differ from a keyword search engine?
It ranks text by semantic and contextual relevance using NLP and deep learning, so it can answer a natural language question with a phrase, a passage or a document.
How does Amazon Kendra help users find relevant information in documents?
Amazon Kendra uses natural language processing and deep learning to rank documents and passages by meaning and context.
A team plans a new enterprise search application on AWS in late 2026. Should it start on Amazon Kendra?
No. Kendra is in maintenance mode and closed to new customers on July 30, 2026. AWS recommends Amazon Bedrock Managed Knowledge Base.