← Back
OpenAI
OpenAI and Amazon launch Stateful Runtime for production AI agents in Bedrock
OpenAI APIOpenAI · releasefeatureapiplatformintegration · openai.com ↗

OpenAI and Amazon Partnership Delivers Production-Ready Agent Runtime

OpenAI and Amazon Web Services have jointly developed a Stateful Runtime Environment that brings production-grade agent capabilities to Amazon Bedrock. This collaboration addresses a critical gap in deploying AI agents at scale: while stateless APIs work well for simple single-turn interactions, real-world business processes require persistent state, multi-step orchestration, and reliable error handling.

What the Stateful Runtime Provides

The runtime runs natively inside AWS environments and is powered by OpenAI models. Key capabilities include:

  • Persistent execution context: Agents maintain memory, history, and workflow state across multiple steps
  • Integrated tool invocation: Seamless execution of tools with automatic state management
  • Error handling and recovery: Long-running tasks can safely resume with preserved context
  • AWS-native governance: Operates within existing security postures, tooling, and compliance frameworks
  • Identity and permission boundaries: Built-in controls for production environments

Simplifies Agent Deployment for Complex Workflows

Development teams no longer need to build custom orchestration layers. Instead of manually managing state storage, tool calls, error handling, and task resumption across disconnected API requests, the runtime handles these concerns automatically. This reduces complexity and accelerates time-to-production for use cases like:

  • Multi-system customer support automation
  • Sales operations workflows
  • Internal IT automation
  • Finance processes with approvals and audits

Availability and Getting Started

The Stateful Runtime in Amazon Bedrock will be available soon. Organizations interested in exploring this capability should contact their OpenAI team or request to be contacted directly for early access and deployment guidance.