All Services

AI / ML

Amazon Kendra

Amazon Kendra is a managed enterprise search service that indexes supported repositories and uses natural-language and relevance models to return answers, document excerpts, FAQs, and ranked results.

Explore pricing models, common use cases, infrastructure support, and the AWS services that commonly work with Amazon Kendra.

Amazon Kendra pricing and cost programs

Pricing model: Search capacity and index usage

On-Demand
Available
Reserved Instances or reserved capacity
Service-specific
Savings Plans
Not applicable
Spot
Not applicable

Billing dimensions: Index hours · Storage units · Query units · GenAI Enterprise Edition usage

Programs and modes: Developer Edition · Enterprise Edition · GenAI Enterprise Edition

Edition and provisioned index capacity determine charges.

Free Tier: Service-specific — verify current offers

Pricing reviewed 2026-07-25. Reviewed against the linked official AWS pricing page. Recheck regional rates and program terms before purchase.

Official AWS pricing

Official AWS sources reviewed 2026-07-21.

Why implement Amazon Kendra?

  • Provides semantic and keyword enterprise search without teams operating a search cluster, crawler fleet, and ranking pipeline.
  • Includes connectors, document metadata, relevance tuning, FAQs, query suggestions, analytics, and access-control features for supported index types and sources.
  • Integrates with enterprise assistants and conversational experiences such as Amazon Q Business and Amazon Lex.

How to implement Amazon Kendra

  1. Inventory sources, formats, owners, classification, source ACLs, identity attributes, freshness targets, expected queries, relevance judgments, and index-type feature requirements.
  2. Create an encrypted index and least-privilege connector roles, store credentials in Secrets Manager, map stable document IDs and useful metadata, and run incremental synchronization with deletion handling.
  3. Pass authenticated user and group context where supported, test allow and deny cases, tune freshness, importance and synonyms from measured judgments, and publish a search interface with click and feedback analytics.

Amazon Kendra best practices

  • Verify access-control support for the selected index type and connector; even when Kendra filters results, the source repository must continue enforcing full-document authorization.
  • Use clean canonical documents, stable IDs, useful titles and metadata, incremental sync, and explicit deletion tests to avoid duplicate, stale, or orphaned results.
  • Evaluate with real queries and graded judgments, inspect zero-result and low-click searches, protect connector credentials, and monitor sync failures, query latency, capacity, and cost.

Amazon Kendra use cases and server impact

  • Enterprise knowledge and policy search
  • Support-agent document discovery
  • Search-backed assistants and conversational bots

Replaces much search indexing, crawling, and ranking infrastructure, while source permissions, content quality, identity mapping, relevance evaluation, and the user experience remain customer-owned.

Official implementation resources

Commonly paired AWS services