A beautiful and performant evolution simulation written in Rust with WebGPU-accelerated graphics and WebAssembly, running entirely in the browser.
Live: evo.tre.systems — needs a WebGPU-capable browser (Chrome / Edge / Brave 113+). It runs a single curated, deterministic world: thousands of gene-carrying creatures evolving, speciating, and hunting in real time.
- Click the canvas to drop a patch of food at the cursor (creatures swarm and graze it down); drag to pan, scroll to zoom, F to frame the swarm, R to reset, H to toggle the control panel (Esc hides it).
- The ⚙ panel tunes the ecosystem (food, breeding, mortality, predators, metabolism), the Food field (how fast patches regrow, how fast they are grazed down, and the floor grazing leaves behind), motion & physics, and Visuals — Glow, Trails, Size, Brightness (instant, cosmetic) plus a Colour lens to recolour the swarm by speed / health / behaviour. Reset / Cull / Bloom act immediately.
- Sound is on by default (it begins on your first interaction, which browsers require) — a fully-synthesised soundscape driven by the live simulation: the chord is the on-screen palette, brightness tracks the swarm's health, the body tracks its size, and dropping food chimes a note pitched and panned to where you clicked. Toggle with M or the ♪ button; volume with ↑ / ↓.
- Rust Nightly:
rustup toolchain install nightly-2026-05-01 - WASM Target:
rustup target add wasm32-unknown-unknown - Node.js & npm: Install Node.js
- WebGPU-capable browser: Chromium-based (Chrome / Edge / Brave) 113+
# Install dependencies and set up environment
npm run setupnpm run dev
# Then open http://localhost:8788Deploy to Cloudflare Pages:
npm run deployFor detailed instructions, see docs/DEPLOYMENT.md.
See the repo layout and code map — kept in one place so they can't drift.
| Command | Description |
|---|---|
npm run dev |
Start local dev server |
npm run build |
Compile to WASM |
npm run deploy |
Build and deploy to Cloudflare Pages |
cargo test |
Run Rust tests |
cargo clippy |
Run linter |
- Architecture — how the code is organized and how a frame is produced
- Diagrams — Graphviz architecture diagrams (system overview, simulation tick)
- Simulation System — genes, movement styles, predation, and statistics
- Deployment — Cloudflare Pages and the SharedArrayBuffer header requirement
- Backlog — ordered next work and known constraints
- Agent Notes — workflow, verification, and architecture rules for agents
MIT License.