procedural-sounds
Procedural interface sounds, with taste. Generated live in the browser from recipes, steered by a hand-curated library and calibrated by ear. No audio files. Beta, open to collaborators.
What is it?
What it is
A procedural audio platform generating interface sounds in-browser via live synthesis from recipes, with export capabilities and curation-driven learning.
Why it exists
To create a collaborative, open-source ecosystem for designing and refining browser-based sound interactions through community-driven training and procedural generation.
Who should use it
sound designersdevelopers interested in generative audioopen-source contributorseducators teaching sound design or programming
Who should avoid it
users requiring pre-recorded audio assetsnon-technical usersthose seeking a polished, production-ready product
How it works
A quick walkthrough in plain English
How procedural-sounds works
Step 1 of 3
You interact with it
Open procedural-sounds, send a request, or connect it to your stack.
Features
Advantages
- Zero audio file storage requirements - everything is synthesized on demand
- Lightweight and fast since no large audio assets need to be loaded
- Highly customizable sounds through recipe-based generation
- Interactive curation workflow allows continuous improvement of sound quality
- Export flexibility with both WAV and JavaScript formats
- Built-in safety features with ear-safety ceilings and loudness calibration
- Comprehensive workbench for advanced users and contributors
- Strong focus on UX with in-context previews and microinteractions
- Active community contribution model with clear pathways for help
Disadvantages
- Early beta status means incomplete features and potential instability
- Requires Node 24 and modern web technologies (Next 16, React 19, Tailwind 4)
- Steep learning curve for the workbench and curation process
- Limited to browser-based synthesis which may not match native audio quality
- Small community (148 stars) may mean slower development and support
- Missing npm package distribution limits integration into other projects
- Local-only dev mode means workbench features unavailable in production builds
- Quality depends heavily on curation effort which requires time and good ears
Installation
native
Node 24 (Next 16, React 19, Tailwind 4). ```bash npm install npm run dev ```
FAQ
How can I contribute to the procedural sounds library?
Contributions are welcome! You can help by curating sounds in the workbench, submitting pull requests for new generators or improvements, or adding to the training data. Check the CONTRIBUTING.md file for specific guidelines.
What is the current state of the project?
The project is in early beta and considered a work in progress. The core functionality (generate, hear in context, export) works, but many features are still being developed. The README emphasizes that it's meant to be actively built upon by the community.
How does the sound generation and learning system work?
The system uses a combination of hand-curated library sounds and procedural generators. When users keep or delete generated sounds in the workbench, this trains the generators by adjusting their probability weights. The library and training data are the same asset that evolves through user interaction.
What are the main technical requirements to run this project?
The project requires Node 24 (or Next 16, React 19, Tailwind 4). You can install dependencies with 'npm install' and run the development server with 'npm run dev'. The workbench features only exist when running locally.
How can I export generated sounds?
You can export sounds as WAV files directly from the interface, or as JavaScript code snippets that can be imported into other projects. This allows for both immediate use and integration into other systems.
What tools are available in the workbench for curation?
The workbench includes tabs for Library browsing, Variations/Creations generation, Editor for manual recipe creation, Calibrate for setting quality thresholds, Dedupe for removing duplicates, and Trash for deleted items. There's also an Atlas view showing generator capabilities.
What are the licensing terms for this project?
The project uses the MIT license. Some initial library sounds came from open source sound packs, with specific licenses detailed in the THIRD-PARTY-NOTICES.md file. All generated content and the procedural system itself are original to this project.
Featured in Videos
YouTube tutorials and walkthroughs for procedural-sounds
Alternatives
Similar projects ranked by category, topics, and text overlap.
