Artificial intelligence news.
Straight from the original announcements.

AWS lets you watch coding agents build web apps live in AgentCore runtime

October 30, 2026, 22:40By the AI News Bites Editorial TeamSource: AWS
AWS lets you watch coding agents build web apps live in AgentCore runtime
Image: AWS

Amazon Web Services has described how Session endpoints in Amazon Bedrock AgentCore runtime let developers open a web app running inside an agent's session and watch it change as a coding agent saves each file. The approach avoids copying the app to another environment and redeploying it for every change.

Key points
  • Session endpoints in Amazon Bedrock AgentCore runtime give each agent session an authenticated URL to open apps running inside it.
  • Developers choose 1 to 5 allowed ports within 8000 to 12000, and every request needs a SigV4 signature or bearer token.
  • In the sample, a Vite server on port 8090 pushes file changes to the browser over WebSocket, so edits show without a refresh.
  • The URL lasts only as long as the session and is meant for review, not for serving apps to users.
  • Session endpoints work with microVM runtimes, not the Instances compute type, and AWS PrivateLink is not supported inbound.

Why the review loop matters

According to AWS, coding agents now build working web apps such as dashboards, product pages and prototypes. The quickest way to review that work is to open the app, use it, ask for a change and check the result. AWS says this works best when the app keeps running where the agent edits it, so each change appears in the browser within seconds. If the app has to be copied out and deployed elsewhere first, every change waits for a deployment and the review slows to that pace.

Amazon Bedrock AgentCore is a platform to build, connect and optimize agents at scale, with a choice of agent framework and foundation model. Its runtime capability runs each agent session in its own isolated microVM. Session endpoints give each session an authenticated URL that forwards each request's HTTP method, path, query string and body, plus WebSocket connections, to the ports the developer allows.

How Session endpoints work

AWS lists six details that shape their use.

  • Ports are chosen by you. You set allowedPorts when calling CreateAgentRuntime: 1 to 5 entries, each a single port or a range within 8000 to 12000, excluding reserved platform ports. Allowing a port does not start a server, so the application must listen on it, bound to 0.0.0.0.
  • The URL comes with the session. The InvokeAgentRuntime response carries it in the X-Amzn-Bedrock-AgentCore-Runtime-Session-Endpoint-Url header.
  • Every request is authenticated. For runtimes using AWS IAM, each HTTP request, HTTP/2 stream and WebSocket handshake carries a Signature Version 4 signature, and the caller needs bedrock-agentcore:InvokeAgentRuntimeProxy. For runtimes with a JWT authorizer, each request carries a bearer token. Knowing the URL does not grant access.
  • The host name selects the port. Clients connect on port 443, and a prefix on the host name, such as 8090, picks the port inside the session.
  • Protocols pass through. HTTP/1.1, HTTP/2 (including server-sent events and gRPC) and WebSocket are carried. The walkthrough uses HTTP/1.1 and WebSocket.
  • It is scoped to one session. The URL lasts only as long as the session.

AWS says the endpoint is meant for building, testing and reviewing what an agent produces. To serve an app to users, it recommends deploying to a hosting service such as AWS Amplify Hosting.

Because a browser cannot add a SigV4 signature or bearer token to the requests a page makes on its own, the Session endpoint client in the bedrock-agentcore SDK can open a local endpoint on the developer's machine. The browser opens http://localhost:8088, and the local endpoint adds credentials to each request and WebSocket handshake before forwarding it.

AWS also distinguishes Session endpoints from runtime endpoints such as DEFAULT. A runtime endpoint controls which version of the agent callers reach, so development and production can run different versions. A Session endpoint shows what the agent runs in its session, works with whichever version the session runs, and stops working when the session ends.

The sample walkthrough

The post walks through a sample with four parts. A coding agent runs in the session on port 8080 and edits a website's files with four tools: list, read, write and edit. The agent starts a Vite development server on port 8090, which serves the site and pushes each saved change to the browser over its hot-reload WebSocket. On the developer's machine, a component called PreviewApp serves a local page with a chat on the left and the live website on the right, sends prompts to InvokeAgentRuntime and opens the local endpoint.

In the example, the user asks the agent to switch the site to a dark color theme, follows its steps as they stream in and sees the site change without a refresh. The agent calls Claude Haiku 5.5 in Amazon Bedrock to decide what to change. A second port, 8081, exposes the agent's /status route through the same Session endpoint. Port 8080 is not in allowedPorts, because InvokeAgentRuntime reaches the agent without going through the Session endpoint.

The sample is in the 14-remote-access folder of the agentcore-samples GitHub repo.

Requirements and limits

AWS says readers need an account in a Region where Session endpoints are available, the relevant IAM permissions, Python 3.10 or later with bedrock-agentcore 1.25.2 or later and boto3 1.43.122 or later, Node.js 22 or later, Git, a Jupyter environment and access to Claude Haiku 5.5 in Amazon Bedrock. The sample deploys the agent as a zip file in Amazon S3 with direct code deployment for Node.js, so Docker and an Amazon ECR repository are not needed.

Session endpoints work with microVM runtimes, not with the Instances compute type. Their URLs are reachable over public HTTPS, and AWS PrivateLink is not supported for inbound connections.

What it means for developers

For teams building with coding agents, the feature lets reviewers see an agent's work in a running app without a separate deployment step for each change. Authentication on every request and the limit to allowed ports keep that access controlled, while the session-bound URL makes it a tool for development and review rather than for serving production traffic.

Related news

AWS shows how to route coding agent requests with NeMo Switchyard
For developersNovember 13, 2026, 23:30

AWS shows how to route coding agent requests with NeMo Switchyard

AWS has published a sample deployment that routes coding agent requests between models on Amazon Bedrock using NVIDIA NeMo Switchyard and LiteLLM. The approach measures…

AWS details four ways to speed up Amazon Bedrock AgentCore sessions
For developersNovember 13, 2026, 23:02

AWS details four ways to speed up Amazon Bedrock AgentCore sessions

AWS has outlined four architecture patterns for reducing session startup delays in Amazon Bedrock AgentCore runtime, including snapshot-based initialization, build hooks…

AWS Says Agents Wrote Its 200+ Service Emulator Without Human Code Review
For developersNovember 12, 2026, 22:52

AWS Says Agents Wrote Its 200+ Service Emulator Without Human Code Review

AWS says autonomous agents wrote and tested every line of AWS Emulator, a free local copy of more than 200 AWS services, while a merge gate controlled which changes…