Sold by: SoftBank Corp.
Deployed on AWS
AGENTIC STAR is an AI agent platform for AWS that automates deployment of an integrated environment with Amazon EKS, Amazon RDS for PostgreSQL, vector search, authentication, observability, and LLM routing through LiteLLM with AWS Bedrock support.
Overview
Image
LoginUI
Image
Chat UI
Image
Admin Cosole
Video
Product video
OVERVIEW
AGENTIC STAR is an AI agent platform delivered as an AWS CloudFormation stack. It automates the deployment and configuration of an integrated environment for building, managing, and running AI agents, including Kubernetes infrastructure, data services, authentication, LLM routing, logging, and monitoring.
AGENTIC STAR is designed for platform engineering, ML, and DevOps teams that want to reduce the manual effort required to integrate and operate multiple AI infrastructure components.
WHY AGENTIC STAR Simplified Infrastructure Integration: Provisions and configures Amazon EKS, databases, vector search, LLM routing, authentication, networking, and observability components through a coordinated deployment. AWS Integration with LLM Flexibility: Routes inference through LiteLLM with AWS Bedrock support. Integration with other model providers available through LiteLLM is subject to configuration and provider compatibility. Integrated Operational Capabilities: Includes components for IAM-based access control, VPC-integrated networking, authentication, monitoring, metrics, and logging. USE CASE EXAMPLE
Example use case: An enterprise knowledge assistant.
A team can deploy AGENTIC STAR and configure an internal support agent that uses company documentation for semantic search and retrieval workflows. Qdrant and PostgreSQL with pgvector provide vector data capabilities, while queries can be routed through AWS Bedrock using LiteLLM.
The Chat UI provides a web interface for employees to interact with the agent. Keycloak supports authentication and team-based access control, while Prometheus, Grafana, Loki, and related monitoring components provide visibility into application and infrastructure health.
KEY FEATURES Automated Deployment: Provisions Amazon EKS, Amazon RDS for PostgreSQL, Amazon S3 storage, VPC networking, and Kubernetes applications through nested AWS CloudFormation templates AI-Ready Data Stack: Includes Qdrant and PostgreSQL with pgvector for vector search, embeddings, and retrieval workflows Chat UI: Provides a web interface for chat, agent configuration, and team management Integrated Observability: Includes Prometheus, Grafana, Loki, Alloy, and Amazon CloudWatch Container Insights Enterprise Security Components: Includes IAM roles, VPC-integrated networking, Keycloak authentication, NGINX Ingress, and cert-manager LLM Routing: Routes inference through LiteLLM with AWS Bedrock support and configurable integration with other compatible model providers Optional Datadog Integration: Provides an optional integration for application performance monitoring and observability WHAT GETS DEPLOYED
AWS Infrastructure:
Amazon EKS cluster with system and user node groups Auto Scaling configuration for EKS node groups Amazon RDS for PostgreSQL with the pgvector extension and Multi-AZ configuration Amazon S3 storage for application data and static content VPC networking with public and private subnets NAT Gateway and Elastic IP IAM roles and policies Bootstrap Amazon EC2 instance for automated Helm installation
Kubernetes Applications and Supporting Components:
Agent Executor AgenticAI Admin AgenticAI ExtAPI Agent Sandbox Controller Chat Chat Auth with Keycloak Keycloak Extensions Qdrant kube-prometheus-stack Prometheus Adapter Loki Alloy Kubernetes Monitoring AgenticAI Observability NGINX Ingress cert-manager ExtAuth Service Gate Services DB Migration Init Platform PVC Lifecycle Manager AgenticAI Datadog Upgrade Controller PRICING
This product uses metered pricing through AWS Marketplace. AWS infrastructure costs, including Amazon EKS, Amazon RDS, Amazon EC2, NAT Gateway, and Amazon S3 usage, are billed separately by AWS based on resource consumption.
PREREQUISITES An AWS account with the IAM permissions required to create the AWS resources used by the product Sufficient Amazon EC2 vCPU service quotas for the Amazon EKS node groups Appropriate service quotas for Amazon EKS, Amazon RDS, Amazon VPC, and Amazon EC2 Access to an LLM endpoint, with AWS Bedrock recommended for use through the LiteLLM gateway EXTERNAL DEPENDENCIES
This product downloads the following external dependencies during deployment:
kubectl from kubernetes.io Helm from helm.sh yq from the mikefarah/yq GitHub repository Helm charts and installer packages from seller-managed object storage Container images from a seller-managed container registry
SoftBank Corp. manages the availability, security, and integrity of artifacts hosted on SoftBank-managed infrastructure. Third-party dependencies are governed by their respective providers.
GETTING STARTED
To learn more about AGENTIC STAR or request a guided deployment walkthrough, contact SoftBank Corp. through the official inquiry form.
Highlights
- Automate the deployment of an integrated AI agent environment through AWS CloudFormation, including Amazon EKS, Amazon RDS for PostgreSQL, storage, networking, authentication, and observability components.
- Support semantic search and retrieval-augmented generation workflows using Qdrant and PostgreSQL with pgvector, with LLM routing through LiteLLM and AWS Bedrock.
- Operate the platform with IAM roles, VPC-integrated networking, Keycloak authentication, NGINX Ingress, cert-manager, Prometheus, Grafana, Loki, and Amazon CloudWatch Container Insights.
Details
Sold by
Delivery method
Delivery option
AGENTIC STAR AMI with CloudFormation (EKS Full Stack)
Latest version
Operating system
AmazonLinux Amazon Linux 2023
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on a fixed subscription cost and actual usage of the product. You pay the same amount each billing period for access, plus an additional amount according to how much you consume. The fixed subscription cost is prorated, so you're only charged for the number of days you've been subscribed. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
- Monthly subscription
- $1,500.00/month
Dimension | Cost/hour |
|---|---|
t3.large | $0.00 |
t3.medium | $0.00 |
m5.large | $0.00 |
m7i.large | $0.00 |
m6i.large | $0.00 |
Dimensions summary
You pay by the hour based on the compute instance you run this platform on. The five options are different AWS EC2 instance types, not feature tiers. Each carries its own hourly rate tied to that instance's compute and memory profile. The t3.medium and t3.large are burstable general-purpose instances. The m5.large, m6i.large, and m7i.large are general-purpose instances from different processor generations. You choose the instance that fits your workload and deploy it to your own environment. Your total cost scales with how many hours you keep the instance running.
Top-of-mind questions for buyers
Each hourly rate is the software charge for running AGENTIC STAR on that EC2 instance type. AWS charges for the underlying compute, storage, and network separately. Your invoice combines both. The instance type you pick sets the software rate and the compute and memory you receive.
Software charges meter running instance-hours. A fully stopped instance stops accruing the hourly software fee. Stopped instances may still incur AWS storage fees for attached volumes, but those are separate from the software charge. You control cost by controlling how long each instance runs.
You can deploy the platform to your own cluster and run multiple instances. Each running instance meters its own hourly software charge based on its type. Your total software cost is the sum of all running instances multiplied by their hours. You can mix instance types across your deployment.
developers.fd.agenticstar.tm.softbank.jp
Helpful?
Vendor refund policy
No refund policy
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
AGENTIC STAR AMI with CloudFormation (EKS Full Stack)
This delivery option uses nested AWS CloudFormation templates to provision the full AGENTIC STAR platform. The entry-point template (main-template.yaml) orchestrates all nested stacks: VPC with NAT Gateway, Amazon EKS cluster with managed node groups, RDS PostgreSQL 16.6 with Multi-AZ HA, S3 buckets, and an ephemeral Setup Helper EC2 instance that installs 23 Kubernetes microservices via Helm.
A placeholder Marketplace EC2 instance (t3.medium) is provisioned in the public subnet to satisfy AWS Marketplace AMI-based product requirements. This instance does not run any application workload.
External dependencies downloaded at deployment time include Helm binaries and charts from seller-managed object storage, and container images from a seller-managed container registry.
CloudFormation Template (CFT)
AWS CloudFormation templates are JSON or YAML-formatted text files that simplify provisioning and management on AWS. The templates describe the service or application architecture you want to deploy, and AWS CloudFormation uses those templates to provision and configure the required services (such as Amazon EC2 instances or Amazon RDS DB instances). The deployed application and associated resources are called a "stack."
Version release notes
Version 2.9.0 release of AGENTIC STAR on AWS Marketplace.
This release upgrades the bundled Helm installer to version 2.9.0, delivering platform improvements through the microservices layer.
Key changes in this release:
- Helm chart renaming: the librechat chart is now agentic-star-front and the librechat-auth chart is now agentic-star-auth; Helm release names and Kubernetes resource names are unchanged, and the upgrade controller migrates existing installations automatically at startup
- Chat UI: per-user default Skills sets (always applied, server-stored, manageable from settings), full-catalog server-side search and pagination for Skills and Playbooks, and numerous UI/UX and multilingual fixes across conversations, library, shared-link management, and mobile layouts with clearer retry paths on data-fetch failures
- Admin console: reworked the dashboard to run without external gateway dependencies, backed by live Prometheus metrics, with a new most-active-organizations view, common unsaved-change guards across forms, and guardrail alert integration with server-to-server routes
- Authentication: OAuth provider support for agent-to-agent (A2A) token acquisition, token reuse with expiry buffering, and fixes for concurrent refresh races and transient blank screens
- Security hardening: unified end-user IP recording in audit logs onto canonical headers, strengthened attachment access control, and enforced fail-closed webhook signature verification
- Platform reliability: enforced replica distribution and quorum PodDisruptionBudget for the vector database, fixed sandbox claim lifecycle to prevent resource leaks, and made the upgrade controller substantially more robust (first-time chart installs, failure-safe state recording, automatic lock release, and automatic catch-up for renamed charts)
- Mail components: new optional inbound mail stack chart (disabled by default) and mail gateway enhancements including raw sending with attachments and a verified, idempotent inbound edge API
- Database: registered migrations for new tables and columns and fixed a schema issue on fresh installations
- Breaking changes: removed the self-service account deletion API, the access-log statistics view, and the temporary-password issuance flow for administrator promotion; old chart names (librechat / librechat-auth) are no longer published as of 2.9.0
All existing components from v2.7.0 (EKS with System/User node groups, RDS PostgreSQL 16.10 with pgvector, Qdrant, Chat UI with Keycloak, Prometheus/Grafana/Loki/Alloy observability, NGINX Ingress, cert-manager, AWS Bedrock integration, optional Datadog APM) remain included.
Additional details
Usage instructions
LAUNCH INSTRUCTIONS
- Subscribe to AGENTIC STAR on AWS Marketplace and click "Continue to Configuration".
- Select "CloudFormation" as the delivery method, choose your AWS Region, and click "Continue to Launch".
- Click "Launch CloudFormation" the AWS CloudFormation console opens with main-template.yaml pre-loaded.
- Fill in the required parameters: Stack name: choose a unique name (e.g., agentic-star-prod) ImageId: pre-filled by AWS Marketplace - do not change EKS node instance types and scaling limits (default: m5.2xlarge, min 2 nodes each group) RDS instance class and credentials (default identifier: marketplace-db) Admin email and password for the platform Primary language (default: ja)
- Acknowledge that CloudFormation will create IAM resources.
- Click "Create stack". Deployment takes approximately 20~30 minutes.
POST-DEPLOYMENT SETUP
After the stack reaches CREATE_COMPLETE:
- Retrieve the EKS cluster name from the CloudFormation Outputs tab.
- Run the kubectl config command shown in Outputs to authenticate with the cluster.
- Verify all 23 microservices are running: kubectl get pods -A
- Access the Chat UI at the URL shown in Outputs.
- Configure your LLM provider through the LiteLLM multiLLM gateway (e.g., AWS Bedrock region) via the platform admin panel.
- Configure Keycloak with your identity provider if required.
For full documentation, refer to the product User Guide.
Resources
Vendor resources
Support
Vendor support
SoftBank Corp. provides support for AGENTIC STAR on AWS through the official inquiry form.
Support Scope:
Assistance with AWS Marketplace deployment and AWS CloudFormation launch issues Guidance for initial configuration and post-deployment setup Documentation for platform administration and operation Support for product updates and security-related notices Troubleshooting for Kubernetes application health and connectivity Guidance on AWS service quotas relevant to Amazon EKS and Amazon RDS resources
How to Get Help:
For deployment issues, initial configuration questions, operational questions, and other product support inquiries, contact SoftBank Corp. using the official inquiry form.
Sizing Guidance:
AGENTIC STAR requires sufficient vCPU service quotas for the Amazon EKS node groups. Ensure that the AWS account has appropriate service quotas for Amazon EKS, Amazon RDS, Amazon VPC, and Amazon EC2 before deployment. Contact support for assistance with capacity planning and service quota requirements. For support inquiries, please contact SoftBank Corp. using the following form: https://www.softbank.jp/biz/contact-us/demand/ai/agentic-star-inquiry/
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
DataRobot's Agent Workforce Platform lets enterprises build, operate, and govern fleets of production AI agents on AWS - and in hybrid, sovereign, and air-gapped environments - with the same governance layer in every one. Model-, framework-, and orchestration-neutral, with forward-deployed DataRobot engineers who build your first mission-critical agents alongside your team.
A governed control plane for deploying AI-built and human-built apps safely inside your own AWS account, with built-in access control, policy enforcement and managed infrastructure so your platform team isn't the bottleneck.
IBM watsonx Orchestrate is an AI agent orchestration platform for AWS to automate HR, sales and support workflows with 700+ integrations.
zeb designs and deploys custom Agentic AI solutions in Amazon Connect, enabling proactive, goal-driven virtual agents that collaborate across systems. Deliver sophisticated self-service experiences and intelligently escalate to live agents when needed.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.