Anima — A GPU laboratory for artificial life

Live demo: https://Nikhil-creat.github.io/anima/

Designed and developed by

★NIKHIL CHARY SRIRAMOJU★

BTech CSE (Final Year)

• GitHub: Nikhil-creat
• LinkedIn: nikhil-chary-sriramoju
• Email: sriramojunikhil66@gmail.com
• Instagram: @nikhil__sriramoju
• Facebook: Profile

Anima simulates continuous cellular automata (Lenia) on the GPU with WebGL2, maps where self-organizing life exists in parameter space, and breeds new genomes with an evolutionary search. Everything runs in the browser. There is no server, no framework and no build step.

Highlights

Capability	Detail
Real-time simulation	Fragment-shader kernel convolution in 32-bit float textures, up to 256 × 256
Interactive dish	Paint or erase cells while the simulation runs
Phase map	Survey of 144 genomes, each classified as extinct, static, saturated or alive
Discover	Ten-generation mutation and selection loop that scores each genome on a full GPU trial
Genome library	Save, load, export and import genomes as JSON; share any genome by URL
Reproducibility	Seeded random number generator, experiment export to CSV
Benchmark	Measures steps per second and billions of kernel taps per second on your GPU
Quality	Dependency-free tested core, CI, automatic deployment

Quick start

Open index.html in a browser that supports WebGL2, or visit the live demo. For local development serve the folder, for example python3 -m http.server.

Tests

node tests/run-node.js        # runs in Node 18+, no dependencies

You can also open tests/index.html in a browser. Both use the same cases in tests/cases.js.

Repository layout

index.html            the lab
assets/css/style.css  styling
assets/js/core.js     pure, tested logic (kernel, seeding, classifier)
assets/js/app.js      WebGL2 rendering and interface
tests/                unit tests for browser and CI
docs/                 architecture and methodology
.github/workflows/    test, then deploy to GitHub Pages

Documentation

• Architecture
• Methodology and limits
• Contributing
• Changelog

Limitations

Anima builds on Lenia, which was created by Bert Wang-Chak Chan. This project's contribution is the interactive GPU implementation, the survey and search tooling, and the engineering around it. The alive-or-not classifier is a heuristic, and outcomes depend on the random seed and the number of trial steps.

Roadmap

• Multi-channel state so several species can compete
• Kernel and growth-function editor
• Larger grids using tiled rendering
• A short study of how often each region of the phase map is alive across many seeds

Citation

See CITATION.cff. GitHub shows a "Cite this repository" button automatically.

License

MIT. See LICENSE.