- AWS reveals the new Amazon rock agent platform
- The launch seeks to get part of the heavy work of the implementation of AI agents
- Includes a selection of tools and services to help build and deploy agents
AWS has revealed a new AI Agent development platform, since it seems to facilitate the construction and implementation of agents easier than ever.
The new Amazon Bedrock Agentcore platform seeks to provide developers everything they need to create and implement advanced AI agents.
Speaking in its AWS Summit New York 2025 event, the company said the launch marks a “change change” to help developers move fun toys agents to something effective and production.
Amazon Bedrock Agentcore
“With the agents they reach a change to the service as software,” said Swami Sivasubramanian, Vice President of Aws Agentic AI, Amazon Web Services, which describes the plans to make AWS the best place to build the most useful AI agents in the world.
“Despite all these frames and open source protocols, bringing agents to production is still too difficult, due to the dynamic and autonomous nature of agents and their effects on production.”
Available in the preview now, with a broader launch that will be available soon, agentcore includes the following services:
Agentcore Runtime: Execution time without safe server built especially to implement and climb AIs and tools
Agent memory: Build agents with the context by eliminating the management of complex memory infrastructure while providing total control over what the AI agent recalls.
Agentcore identity: Securely access AWS services and third -party tools on behalf of users or acting with prior authorization
Agentcore Gateway: Build, implement and discover agents in millions of connections – automatically become MCP compatible tools without administering integrations
Agent Code Interpreter: Allow IA agents to write and execute code safely, improving precision to solve complex end -to -end tasks, including JavaScript and Python
Agentcore browser tool: Quick and safe browser execution time based on the cloud to allow AI agents to interact with scale websites, including live display for resolution and audit
Observability of agentcore – trace, purification and monitoring the performance of AI agents in production environments
Users can choose what tools they want to use from the platform, which allows them to climb and implement when they are ready.
“It is a tectonic change in some dimensions,” Sivasubramanian added. “It reduces the way in which software is built. It also presents a lot of new challenges to implement and operate it, and potentially in a more shocking way, change how software interacts with the world and how we interact with software.”
