![[Announcement] Amazon Bedrock AgentCore is now generally available (GA)!](https://images.ctfassets.net/ct0aopd36mqt/7qr9SuOUauNHt4mdfTe2zu/8f7d8575eed91c386015d09e022e604a/AgentCore.png?w=3840&fm=webp)
[Announcement] Amazon Bedrock AgentCore is now generally available (GA)!
This page has been translated by machine translation. View original
Hello, I'm Jinno from the Consulting Department, who loves supermarkets.
Finally, the general availability of Amazon Bedrock AgentCore (hereinafter AgentCore) has been announced! Hooray!!
AgentCore was released as a public preview in July 2025, and now it's already GA...!!
What is Amazon Bedrock AgentCore
Amazon Bedrock AgentCore is an optimal managed service for deploying and operating AI agents.
It offers a variety of managed features that are convenient for developing AI agents.
- Runtime
- Hosting functionality
- Identity
- Authentication functionality
- Gateway
- Tool-ification functionality for external processing
- Memory
- Memory functionality
- Built in tools
- Code Interpreter: Code execution environment
- Browser: Browser execution environment
- Observability
- Tracing and visualization of AI agents
New regions and features added with GA
Added regions
While AgentCore was available in 4 regions during the public preview, with GA it has expanded to 9 regions!
The regions that were available during the preview are as follows:
- US East (N. Virginia)
- US West (Oregon)
- Asia Pacific (Sydney)
- Europe (Frankfurt)
And here are the new regions added with GA!
- US East (Ohio)
- Asia Pacific (Mumbai)
- Asia Pacific (Tokyo)
- Asia Pacific (Singapore)
- Europe (Ireland)
The Tokyo region has been added!! That's great news!
Newly added features
There are also features that were added at the time of GA.
MCP Server added as a Gateway target
Connection functionality to MCP (Model Context Protocol) servers via AgentCore Gateway.
MCP is a protocol that standardizes communication between AI agents and tools. With AgentCore Gateway supporting this, you can now connect from Runtime to existing MCP servers via Gateway.
You specify the remote MCP Server endpoint and select an authentication scheme. Authentication schemes include OAuth or no authentication, with OAuth being recommended.
When using tool search via Gateway, the appeal seems to be that you can add MCP servers through Gateway without modifying the agent code. I haven't fully organized my thoughts on the appeal yet, so I'd like to try it out and share my impressions!
A2A protocol support
Another important new feature is support for the A2A (Agent-to-Agent) protocol!
The A2A protocol is a standard protocol for AI agents to communicate with each other and work collaboratively. This allows multiple agents to work together to handle complex tasks. Please refer to the official documentation for implementation details!
I'd like to try this and write a blog about it!
By the way, AWS has also published an official A2A sample implementation which should be helpful when trying it out!
Gen AI Dashboard GA
The convenient Gen AI Dashboard, which is an Observability feature, has also reached GA!
AgentCore Feature Blog Introduction
Please refer to the blog I've summarized below.
I've also written blogs about each feature, so please take a look.
Runtime
Identity
Gateway
Observability
Built in tools
Code Interpreter
Browser
Pricing
Please refer to the official page below.
Conclusion
Finally, AgentCore has reached GA...!!!
It's a service with convenient features for creating AI agents on AWS. Let's use it actively and create useful AI agents! I'll continue to share knowledge abundantly!
I'll also try out the newly added features like A2A support and Gateway MCP support and write blogs about them!
I hope this article has been helpful! Thank you for reading until the end!!
