LinuxCommandLibrary

aws-kendra

Manage ML-powered enterprise search indexes and data sources.

TLDR

Create a new Kendra index

$ aws kendra create-index --name [my-index] --role-arn [arn:aws:iam::account:role/kendra-role]
copy
List all indexes in the account
$ aws kendra list-indices
copy
Submit a search query
$ aws kendra query --index-id [index-id] --query-text "[search terms]"
copy
Add documents to an index
$ aws kendra batch-put-document --index-id [index-id] --documents file://[documents.json]
copy
Create a data source connector
$ aws kendra create-data-source --index-id [index-id] --name [s3-source] --type S3 --configuration file://[s3-config.json] --role-arn [arn:aws:iam::account:role/kendra-role]
copy
Start a data source sync
$ aws kendra start-data-source-sync-job --index-id [index-id] --id [data-source-id]
copy
Create a FAQ for direct answers
$ aws kendra create-faq --index-id [index-id] --name [my-faq] --s3-path Bucket=[bucket],Key=[faq.csv] --role-arn [arn:aws:iam::account:role/kendra-role]
copy

SYNOPSIS

aws kendra command [options]

DESCRIPTION

aws kendra is the AWS CLI interface for Amazon Kendra, an intelligent enterprise search service powered by machine learning. Kendra provides natural language search capabilities across documents, FAQs, and other content sources.
Kendra indexes documents from various data sources including S3, SharePoint, Salesforce, ServiceNow, and databases. It understands natural language queries and returns relevant passages with highlighted answers, supporting semantic search beyond keyword matching.

COMMANDS

create-index

Create a new search index
delete-index
Delete an index
list-indices
List all indexes
query
Submit a search query
batch-put-document
Add documents to an index
batch-delete-document
Remove documents from an index
create-data-source
Create a connector to external content
start-data-source-sync-job
Trigger document synchronization
create-faq
Add FAQ content for direct answers
get-query-suggestions
Get autocomplete suggestions

CAVEATS

Index creation takes 15-30 minutes. Document ingestion costs apply per document scanned. Query capacity units (QCUs) determine concurrent query limits. Large document processing can be slow; use data sources for bulk content. Enterprise edition required for advanced features.

HISTORY

Amazon Kendra launched in December 2019 as an ML-powered enterprise search service. Query suggestions were added in 2021, and featured results for promoted content in 2022. The service has expanded data source connectors to support over 14 content repositories.

SEE ALSO

> TERMINAL_GEAR

Curated for the Linux community

Copied to clipboard

> TERMINAL_GEAR

Curated for the Linux community