How to use Claude to prototype apps in a live host
A Claude Code CLI agent uses a host-issued token to create or modify apps in a prototyping lab. A reload rebuilds backend and UI components, after which the agent tests through browser or API and the author inspects the running result.
news.ycombinator.com
Prompt to try
Editorial adaptation, not the source author’s exact prompt. Replace bracketed inputs and review tool permissions first.
Using [host token] and [lab procedures], create or modify [app] with [backend entrypoint] and [UI components]. Respect [host API and multi-worker constraints]. Run [reload command], test [behaviors] through the browser or API, and present the running interface for my inspection.
Reach people exploring AI agents
See placements →What happened
The host reloads the changed backend and rebuilds UI widgets while other apps stay up; the agent tests the result and the author inspects it in the frontend.
Reported by elenaviter on Hacker NewsHow the agent did it
- Set up the prototyping lab and a Claude Code plugin for operations and app authoring.
- Issued the agent a token in the host application.
- Had the agent add or change apps and run the host reload command.
Where a person stepped in
The author looks at the running result in the lab frontend and can interact directly with app UI components.
A person approved each step before the agent continued.
The author set up an app prototyping lab and a Claude Code plugin containing procedures for operating the lab and authoring apps. The coding agent works from the CLI rather than as an API-paid chat resident inside the application. The lab hosts and runs the generated apps, which combine a Python backend with UI components; the author also added support for Node sidecars.
To authorize changes, the author issues the agent a token in the host application. The agent can add or remove an app, change backend or frontend code and configuration, then run a reload command. The host reloads the backend, rebuilds UI widgets and serves them while other hosted apps remain running. The agent then tests its work through the browser or API.
The author inspects the running result through the lab's frontend, where app UI components can be used directly. Authentication is configured at host level, and generated code belongs to its own app and is exposed through that app's APIs. The author notes that apps must account for a multi-worker, multi-machine host. Who it's for: developers prototyping hosted applications who want an agent to implement and test changes while they inspect the working interface.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · elenaviter
Frequently asked questions
How did the agent access the host?
The author issued a token in the host app. The CLI agent could then add or remove apps, modify their code or configuration and run a reload command.
How were changes checked?
After reload, the agent tested what it built through the browser or API. The author looked at the running result in the lab frontend.
More details
Remaining steps
- Had the agent test the running app through browser or API.
- Inspected the running result in the lab frontend.
About this account
The CLI agent could create, remove or modify hosted apps and reload them, then test the result through browser or API.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=49374992. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: elenaviter on Hacker News
Source published 20 Aug 2026
Source checked 14 Oct 2026
Tools and topics
Claude Code · Claude Code plugin · Python · React · Node · browser/API tests
prototyping · hosted apps · live reload
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.