Introducing the subsequent technology of Amazon OpenSearch Serverless for constructing your agentic AI functions

As we speak, we’re saying the subsequent technology of Amazon OpenSearch Serverless, a completely managed search and vector engine designed for patrons constructing AI brokers. The following technology of OpenSearch Serverless scales from zero to 1000’s of requests per second and again to zero when idle, providing as much as 60% value financial savings in comparison with the price of OpenSearch Service clusters provisioned for peak capability.

The following technology of OpenSearch Serverless creates sources in seconds and scales capability as much as 20 instances quicker than the earlier technology. With on the spot useful resource creation and native integrations with AI growth platforms like Vercel and Kiro, you possibly can deploy production-ready search and vector backends on your AI brokers in minutes with out managing infrastructure.

The following technology of OpenSearch Serverless in motion

To get began with the subsequent technology of OpenSearch Serverless, select Create assortment within the Serverless menu within the Amazon OpenSearch Service console.

Create NextGen assortment with on the spot auto scaling and scale-to-zero for value optimization. At launch, we assist full-text search and vector search just for the gathering kind. If you wish to use the prevailing OpenSearch Serverless infrastructure, select Change to Traditional.

Select Categorical create, the quickest approach to create assortment. No configuration is required—the default settings and matching safety insurance policies are utilized mechanically. Some configuration choices will be modified later.

If you select Create assortment, OpenSearch Serverless will provision sources in seconds.

You may also create a set of OpenSearch Serverless with AWS Command Line Interface (AWS CLI) or AWS SDKs. Here’s a pattern CLI command to create a set group.

aws opensearchserverless create-collection-group 
    --name channy-nextgen-group 
    --standby-replicas ENABLED 
    --generation NEXTGEN 
    --description "My NextGen assortment group" 
    --capacity-limits '{
        "maxIndexingCapacityInOCU": 96,
        "maxSearchCapacityInOCU": 96,
        "minIndexingCapacityInOCU": 0,
        "minSearchCapacityInOCU": 0
    }' 
    --region "us-east-1"

Now, you possibly can create a set that inherits the technology from its mum or dad assortment group. Supported assortment varieties: SEARCH and VECTORSEARCH.

aws opensearchserverless create-collection 
    --name channy-nextgen-collection 
    --type SEARCH 
    --collection-group-name channy-nextgen-group 
    --standby-replicas ENABLED 
    --description "My assortment in NextGen group" 
    --region "us-east-1"

To study extra about managing the subsequent technology of OpenSearch Serverless, go to the Amazon OpenSearch Serverless documentation.

Constructing your brokers quicker with OpenSearch Serverless

To assist constructing production-ready agent functions in Vercel, now you can create a brand new OpenSearch assortment or join your current OpenSearch Serverless assortment throughout the Vercel console. Create a search backend in seconds and add options on-demand as your software grows. To study extra, go to AWS for Vercel.

You may go from concept to working prototype in minutes utilizing Claude Code, Cursor, and Kiro. OpenSearch Agent Expertise present a repository of abilities that convey OpenSearch intelligence immediately into your agent. Every ability encapsulates area information, finest practices, and multi-step execution logic for a particular workflow–so your agent not solely will get outcomes, however understands how they had been achieved. You may also use the OpenSearch Launchpad in Kiro Powers to speed up search functions with guided, end-to-end structure planning.

Now accessible

The following technology of Amazon OpenSearch Serverless is mostly accessible at present and is offered in all AWS business Areas the place Amazon OpenSearch Serverless is at present accessible.

The following technology of OpenSearch Serverless costs for the compute you utilize in OpenSearch Compute Items (OCUs) for indexing, search, and GPU acceleration. You’re charged individually for storage in GB-month. For extra data, see Amazon OpenSearch Service Pricing.

Give it a try to ship suggestions to the AWS re:Publish for Amazon OpenSearch Service or by means of your typical AWS Help contacts.

— Channy

Up to date on Could, 29, 2026 — Fastened the default worth (96) of most indexing/search capability within the CLI command. It is best to use the variety of binary sequence to set these values.