AWS Debuts Bedrock AgentCore to Simplify Secure AI Agent Deployment

At the AWS Summit in New York, Amazon Web Services unveiled Amazon Bedrock AgentCore (Preview)—a modular, enterprise-grade platform designed to help developers deploy, manage, and scale AI agents securely and efficiently.

Earlier this week, it was rumoured that AWS is poised to unveil its AI Agent Marketplace on July 15 at the AWS Summit in New York City, with Anthropic—creator of the Claude AI models—lined up as a launch partner.

The platform supports any framework, model, and protocol, enabling production-ready deployment of agentic AI.

AgentCore comprises seven key services:

• AgentCore Runtime: A serverless environment with session isolation, low latency, and support for long-running workloads.
• AgentCore Memory: Handles short- and long-term context management.
• AgentCore Identity: Manages secure access to AWS and third-party services like GitHub and Slack.
• AgentCore Observability: Offers tracing, debugging tools, and OpenTelemetry-compatible dashboards.
• AgentCore Gateway: Converts APIs and Lambda functions into agent-ready tools via Model Context Protocol (MCP).
• AgentCore Browser: A secure, headless browser runtime for web-based workflows.
• AgentCore Code Interpreter: Sandboxed code execution across multiple languages.

“AgentCore eliminates tedious infrastructure work and operational complexity so development teams can bring agentic solutions to market faster,” AWS said in a blog post.

The platform aims to bridge the gap between proof-of-concept and scalable, secure deployments.

According to the company, early adopters such as Itaú Unibanco, Innova ccer, Epsilon, and Box are already leveraging AgentCore to build secure, compliant AI agents in verticals such as banking, healthcare, marketing, and content management.

AgentCore is available in preview across regions including US East (N. Virginia), US West (Oregon), Asia Pacific (Sydney), and Europe (Frankfurt).