Skip to main content
A live demo is available at https://jkh2.github.io/Primordial-Sim/ — no installation required.
Primordial is a single self-contained HTML file (index.html). There is no build step, no npm, no dependencies, and no server required. Hosting it is as simple as serving one file.

​
GitHub Pages

1

Fork or clone the repository

Fork jkh2/Primordial-Sim on GitHub, or clone it to your own account:
2

Enable GitHub Pages

In your repository, go to Settings → Pages. Under Source, select the main branch and click Save.GitHub will begin building and deploying the site automatically. This usually takes under a minute.
3

Visit your deployment

Once Pages is enabled, your simulation will be live at:
Replace [username] with your GitHub username.

​
Running Locally

No server is needed. Open index.html directly in any modern browser:
The full simulation runs immediately — organisms, physics, WebGL rendering, sound, and all UI panels. Nothing is fetched from a server at runtime.

​
Static File Hosts

Any static file host works. Drop index.html into Netlify, Vercel, Amazon S3, Cloudflare Pages, or any web server and Primordial will run without modification. No configuration files, no build output directories — just one file.

​
Browser Requirements

Primordial requires any modern browser with WebGL support. No WebGL2 is required.

Supported Browsers

  • Chrome
  • Firefox
  • Safari
  • Edge

Not Required

  • WebGL 2
  • Browser extensions
  • Plugins
  • A network connection (for the simulation itself)
If WebGL is unavailable, the page will display a message: WebGL not supported.

​
AI Lab Partner & Offline Use

The simulation itself — organisms, evolution, physics, sound — runs fully offline. No network requests are made during normal operation. The AI Lab Partner (right-side panel) makes API calls to the AI provider you configure:
  • Cloud providers (OpenAI, Anthropic, xAI): require internet access and an API key.
  • Local models (Ollama, LM Studio, vLLM, or any OpenAI-compatible server): work fully offline. Set the provider to Custom / Local and point the endpoint URL at your local server (e.g. http://localhost:1234/v1/chat/completions). No API key is required.
API keys are stored in your browser’s localStorage only — they are never sent to any server other than the AI provider you select.