How to use Claude to build browser music generators
Fable in Claude Code produced two browser music generators from different prompting approaches, inspected them in Chrome, and adjusted their files without assessing the audio.
news.ycombinator.com
Original prompt
Shared by the source author. See source notes for context.
Write a browser-based synthesizer for me. The synthesizer should automatically create interesting polyphonic music in which the various voices play off against each other in both harmony and contrast. The controls will affect the tone, rhythmic patterns, number of voices, complexity and randomness of the melodies, and other features. The controlled features should be original—not just standard synthesizer functions—and encourage creative explorations even by naive users.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source poster / publisher
- tkgally
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Two browser-based music generators were produced in one-shot attempts. Fable adjusted their files after browser inspection but could not assess the resulting sounds.
- Human involvement
- The author compared the two attempts without declaring a winner; the post does not describe a detailed audio review.
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 Fable in chat to expand a metaprompt into a full synthesizer prompt.
- Passed the expanded prompt to Fable in Claude Code.
- Supplied a shorter synthesizer prompt directly for a second generator.
Tkgally used Fable in Claude Code to create two browser-based music generators. For the first, they asked Fable in chat to write a full prompt using an earlier metaprompt, then passed that expanded prompt into Claude Code. For the second, they supplied a short request directly to Fable in Claude Code.
The direct request described a synthesizer that automatically creates polyphonic music, with voices interacting through harmony and contrast. It asked for controls affecting tone, rhythmic patterns, voice count, melodic complexity, randomness, and other original features suitable for exploration by inexperienced users. Those were requested properties; the post does not provide a feature-by-feature assessment of their implementation.
The author described both attempts as completely one-shot. During each attempt, Fable opened the files in Chrome, inspected and adjusted the page layout, monitored browser events for errors, and changed the file. It could not listen to or assess the generated sounds. The author did not declare either prompting approach better, noting that comparison would require more testing and subjective judgments.
Who it's for: People prototyping interactive music tools who can personally assess the audio.
Prompt
Original prompt
Source: Hacker News · tkgally
Frequently asked questions
Did the longer prompt produce a better synthesizer?
The author did not make that claim. They said determining which result was better would require considerably more testing and would involve subjective judgments.
What could Fable check automatically?
It opened files in Chrome, inspected and adjusted layout, monitored browser events for errors, and modified the files. It could not listen to or assess the sounds.
More details
Remaining steps
- Fable opened the files in Chrome, inspected layout and events, and adjusted each file.
About this account
Both generators were described as one-shot attempts, including autonomous browser inspection and file adjustments, but no automated audio assessment.
Platform: Hacker News. Attribution: https://news.ycombinator.com/item?id=48755889. Prompt: Original (quoted from source). Sourced by the Agents In Use pipeline.
Credit: tkgally on Hacker News
Source published
Added to Agents In Use
Source checked
Tools and topics
Claude Code · Fable chat · Chrome
Music · Synthesizer · Browser apps · Prototyping
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.