SakanaAI

SakanaAI / digital-ecosystem

Public

Interactive multi-agent NCA ecosystem simulation

pub.sakana.aidigital-ecosystem
40
7
100% credibility
Found Apr 19, 2026 at 40 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

A browser-based interactive simulation of competing neural cellular automata species evolving on a 2D grid, with tools to draw, tune parameters, record, and save states.

How It Works

1
🔍 Discover the digital ecosystem

You stumble upon a fascinating blog post about virtual life forms battling for space in a colorful grid world.

2
🌐 Open in your browser

Just click the live demo link or download and open the simple webpage—no setup needed, it works instantly.

3
🚀 Load a ready-made world

Hit 'Load Showcase' and watch as colorful species burst to life, spreading, fighting, and evolving right before your eyes.

4
👆 Draw and shape the world

Grab the drawing tools to build walls, plant new species, or erase areas, steering the drama as territories clash.

5
⚙️ Tune the rules of life

Slide controls to tweak competition, growth speed, or learning rates, discovering wild new behaviors.

6
💾 Capture your creation

Record a video of the action or save a snapshot to revisit or share your unique ecosystem later.

✨ Share thriving worlds

Your custom digital habitats come alive with emergent patterns—endless fun experimenting with AI life!

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 40 to 40 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is digital-ecosystem?

Digital-ecosystem is a browser-based simulation of competing neural cellular automata species battling for territory on a 2D grid, modeling a digital ecosystem where agents evolve strategies in real time through gradient descent. Built in JavaScript with TensorFlow.js and WebGL2, it lets you load pre-trained showcases, draw walls or seed species, and tweak over 20 parameters via an interactive control panel—no build step or server needed, just open index.html. It's a hands-on way to explore emergent behaviors like coexistence and chaos transitions in multi-agent systems.

Why is it gaining traction?

It stands out with differentiable competition mechanics that allow soft outcomes beyond winner-take-all, plus features like video recording, timeline visualizations of populations and loss, and checkpoint saves for branching experiments. Developers dig the live intervention tools and seeded reproducibility, turning abstract NCA research into an engaging digital ecosystem model you can poke and record instantly. With a companion blog unpacking dynamics like optimizer impacts, it's hooking folks prototyping interactive life sims.

Who should use this?

AI researchers studying emergent cooperation or edge-of-chaos in neural cellular automata will love the real-time training and metric tracking. Game developers building procedural ecosystems or Marvel Rivals-style agent battles get a ready canvas for custom species. Educators and hobbyists wanting interactive GitHub Pages demos of digital ecosystems find it perfect for classrooms or interactive readmes.

Verdict

Try it for NCA experimentation—solid for demos and prototyping, with intuitive controls and exportable .petri bundles. At 40 stars and 1.0% credibility, it's early-stage with thin docs and no tests, so expect tweaks for production use.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.