what is loom fr?

Loom for AWS is a management platform for building, deploying, and running AI agents on Amazon Bedrock AgentCore Runtime and AWS Strands Agents. It gives teams a single web interface to handle agents, their memory, connected tools, and agent-to-agent integrations, along with login and permission controls, instead of managing each of those pieces separately through raw AWS services. From the interface, you can deploy new agents or bring in existing ones, including a no-code option that lets you configure model settings and built-in tools like a code interpreter or browser without writing code. Loom also supports agents that need private network access, with configuration for connecting to internal resources inside a company's virtual network. It can route an agent's requests through Amazon Bedrock or through an alternate AI provider connected via a proxy, and it tracks token usage and estimated cost per agent so teams can see what their agents are spending. The platform includes a memory system for agents to remember past interactions using different strategies such as summarizing conversations or tracking user preferences, plus tools to register and manage MCP servers and agent-to-agent connections with their own authentication and access rules. An optional governance feature lets administrators require agents and tools to be reviewed and approved before they can be used in production. Security features include login through Amazon Cognito, support for outside identity providers like Microsoft Entra ID or Okta, and a two-part group system that controls both what a user sees in the interface and what resources they can access. There is also an approval workflow for actions that need human review before an agent is allowed to proceed. An admin dashboard tracks usage, logins, and actions across the platform for oversight. The project is written primarily in Python for its backend, built with FastAPI and boto3, with a React and TypeScript frontend. It is intended for organizations, specifically platform teams and administrators managing AI agents at scale on AWS, rather than individual hobby developers.