How to use Claude to build a browser jam app
The author reports a 1,529-line single HTML jam app with nine tracks, sequencer and keyjazz modes, mobile multi-touch and WebRTC audio, tested with two browsers and a phone.
news.ycombinator.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Using [AIMEAT node documentation], build a single HTML multi-user band jam app for [node]. Include nine tracks, step sequencing, live keyjazz, virtual drums and pitched instruments, Live/Edit/Rec modes, mobile multi-touch and WebRTC peer-to-peer audio. Iterate with me and prepare it for a [devices] jam test.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- hamuf
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- The author reports a 1,529-line single HTML jam app with nine tracks, sequencer and keyjazz modes, mobile multi-touch and WebRTC audio, tested with two browsers and a phone.
- Human involvement
- The author participated in iteration rounds and tested the application by jamming locally with browsers and a phone.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Asked Opus to build a multi-user band jam app on an AIMEAT node.
- Iterated on the app over several rounds in one evening.
- Ran the resulting single HTML file on AIMEAT.
AIMEAT's author asked Opus to build a multi-user band jam application on their own node. They describe several rounds of iteration over one evening, followed by a working application consisting of a single HTML file with 1,529 lines. It drops into AIMEAT and runs in the browser.
The reported app has nine tracks, a ProTracker-style step sequencer, live keyjazz, virtual drums and pitched instruments. It supports Live, Edit and Rec modes, mobile multi-touch and WebRTC peer-to-peer audio. The author tested the result by jamming with two browsers and a phone in the same room.
The surrounding platform supplies building blocks including identity, memory, actions, a work queue, a token ledger, boards, federation and observability. The author explains that AIMEAT apps are browser-based single HTML files and that AI chats can write apps from the platform's prompts. For this jam application, however, the post gives the initial request in paraphrase rather than reproducing the complete prompt.
The author also has a friend's node peered with their own, describing a federation of two. That platform connection is reported separately from the local jam test; the post does not claim a remote cross-node music session.
Who it's for: Musicians and developers who want to prototype a shared browser music tool on an AIMEAT node.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: Hacker News · hamuf
Frequently asked questions
What features did the jam application include?
The author lists nine tracks, a ProTracker-style sequencer, live keyjazz, virtual drums and pitched instruments, Live/Edit/Rec modes, mobile multi-touch and WebRTC P2P audio.
How did the author try the generated application?
After several iteration rounds in one evening, the author jammed with two browsers and a phone in the same room. The result was a single HTML file running on their AIMEAT node.
More details
Remaining steps
- Jammed with two browsers and a phone in the same room.
About this account
Opus built the browser application through several iterations using the AIMEAT node as its target environment.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=48023146. Prompt: Suggested (written by Agents In Use, not the author's). Sourced by the Agents In Use pipeline.
Credit: hamuf on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude Opus · AIMEAT · HTML · WebRTC
Music app · AIMEAT · WebRTC
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.