AWS Launches Open-Source 'Loom' Platform for Secure Enterprise AI Agent Deployment

Amazon Web Services (AWS) on Thursday released Loom for AWS, an open-source platform designed to help enterprises build, deploy, and manage AI agents with robust security and governance controls. The new offering targets platform engineering teams and aims to address critical enterprise requirements for deploying AI technologies safely and at scale.

A Focus on Governance and Security

Loom provides a framework for managing the entire lifecycle of AI agents, with a strong emphasis on security and compliance. According to AWS, the platform uses configuration-driven methods rather than runtime code generation, enhancing predictability and control. Credentials and other sensitive data are managed through AWS Secrets Manager.

Key governance features baked into the platform include:

• Automated Tagging: Loom enforces three required tags on all deployed resources and supports custom tagging for better cost attribution and governance.
• Access Control: It implements a two-dimensional system combining role-based and attribute-based controls to granularly restrict user permissions.
• Approval Workflows: The platform includes human-in-the-loop approval processes for sensitive actions, ensuring manual oversight where necessary.

Integration with the AWS Ecosystem

Ad

The platform is built to integrate deeply with existing AWS services. It works with Amazon Bedrock AgentCore and AWS Strands Agents to manage agent lifecycles. For deployment, Loom supports both low-code options using pre-written Python agents and no-code deployments through AgentCore's managed system, catering to different levels of technical expertise.

Loom also connects with the AWS Agent Registry, which is currently in public preview. This integration allows enterprises to manage agent and tool records while enforcing governance reviews before any agent is deployed into a production environment. The system uses OAuth2 protocols to ensure user identity is securely propagated through agent request chains.

Market Context and Availability

The release of Loom signifies a move by AWS to provide the foundational tooling that large organizations need to adopt generative AI agents without compromising on security or internal policies. By open-sourcing the project, AWS encourages community contributions and broader adoption.

The project is immediately available through AWS Labs on GitHub. AWS states that it is intended for platform engineering teams tasked with building applications on top of fully managed AWS services, positioning it as a key component for enterprise-grade AI infrastructure.