TLDR: Amazon Web Services (AWS) has rolled out a new capability for Amazon Bedrock AgentCore Runtime, allowing users to configure custom domain names for their AI agents. This enhancement aims to provide a more professional, branded, and seamless user experience by transforming default AWS endpoints into user-friendly URLs, leveraging services like Amazon CloudFront, Route 53, and AWS Certificate Manager.
Amazon Web Services (AWS) announced on August 29, 2025, the introduction of custom domain name support for Amazon Bedrock AgentCore Runtime agents. This new feature addresses a common customer need to move beyond generic AWS endpoints, such as `https://bedrock-agentcore.{region}.amazonaws.com/runtimes/{EncodedAgentARN}/invocations`, to more recognizable and branded URLs like `https://agent.yourcompany.com` .
The primary motivation behind this update is to enhance the user experience and branding for client-side applications, including web browsers and mobile apps, that interact with AgentCore Runtime agents. Custom domains offer improved brand recognition, a more professional appearance, and a seamless integration experience, especially when using Open Authorization (OAuth) authentication where applications make direct HTTPS requests to the service endpoint .
The technical solution for implementing custom domains involves using Amazon CloudFront as a reverse proxy. CloudFront is configured to transform requests from the custom domain into Amazon Bedrock AgentCore Runtime API calls. This setup is complemented by Amazon Route 53 for DNS resolution and AWS Certificate Manager (ACM) for managing SSL/TLS certificates, ensuring secure and scalable custom domain functionality .
Amazon Bedrock AgentCore Runtime, currently in preview, is a secure, serverless, and purpose-built hosting environment designed for deploying and running AI agents or tools. It is framework-agnostic, supporting popular frameworks like LangGraph, CrewAI, and Strands Agents, as well as custom-built solutions. Key features of AgentCore Runtime include extended execution times of up to 8 hours for complex reasoning tasks, session isolation through individual microVMs for enhanced security, built-in authentication (supporting IAM SigV4 and OAuth 2.0), and a consumption-based pricing model .
This capability is part of the broader Amazon Bedrock AgentCore suite of services, which aims to simplify the development and deployment of enterprise-grade AI agents. Other components of this suite include AgentCore Identity for managing agent identities and credentials, AgentCore Gateway for integrating existing APIs and Lambda functions as agent tools, AgentCore Observability for tracing and debugging, AgentCore Memory for short-term and long-term state persistence, AgentCore Code Interpreter for secure code execution, and AgentCore Browser for web interaction .
Also Read:
- AWS Propels Autonomous Operations with Advanced Agentic AI Capabilities
- Availity Revolutionizes Healthcare Operations with AWS Generative AI Integration
Swami Sivasubramanian, AWS’s VP for agentic AI, highlighted the significance of AgentCore, stating that it represents a ‘step change’ for securely transitioning agents from proof-of-concept to production. He emphasized that it ‘provides tools and capabilities to make agents more effective and more capable, purpose-built infra to securely scale agents and provide the controls to enable you to build trustworthy agents’ . This latest update further solidifies AWS’s commitment to providing robust infrastructure for AI agent development and deployment.
