0xSanei

0xSanei / darwinia

Public

The Self-Evolving Agent Ecosystem — Trading agents that evolve through Darwinian selection and adversarial self-play

colab.research.google.comgithub0xSaneidarwiniablobmainnotebooksquickstart.ipynb adversarial-testing ai-agents autonomous-agents backtesting bitcoin
46
4
100% credibility
Found Apr 07, 2026 at 46 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Darwinia is an open-source Python tool for evolving adaptive cryptocurrency trading agents through genetic algorithms, adversarial testing, and pattern discovery using historical market data.

How It Works

1
🧬 Discover Darwinia

You find Darwinia, a fun tool where trading bots evolve their own smart strategies just like animals in nature, no hand-coding needed.

2
💻 Set it up easily

Download the files and install with one simple command, everything ready in seconds on your computer.

3
🚀 Start evolving agents

Hit go and watch dozens of little trading agents compete, survive, and get better at making money over many exciting generations.

4
⚔️ Test the champion

Challenge your best agent with sneaky market traps like fake surges or sudden crashes to prove it's tough.

5
📊 View the dashboard

Open a beautiful screen showing family trees of winners, battle histories, and secret patterns they discovered.

🎉 Trading strategies evolved!

Celebrate having adaptive bots that handle real market chaos, ready for you to study and use ideas from.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 46 to 46 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 darwinia?

Darwinia builds self-evolving trading agents using Darwinian selection and adversarial self-play on real BTC data. Feed it crypto candles, and it spawns populations with DNA encoding signals, thresholds, and personalities; weak agents die in arenas against rug pulls or fake breakouts, while survivors breed fitter offspring. Python-based with numpy/pandas, it offers CLI commands like `evolve`, `arena`, `fetch` from Binance/CoinGecko, and a Streamlit dashboard—all local, no APIs needed.

Why is it gaining traction?

Unlike static backtesters, Darwinia pits agents against targeted attacks that read their DNA, sparking an arms race that uncovers novel patterns automatically. JSON outputs feed agent pipelines (OpenClaw/Claude skills), walk-forward validation catches overfitting, and macro regimes reward adaptive sizing. It's an awesome self-evolving agents github pick for darwinian evolution without reinventing genetic algos.

Who should use this?

Crypto traders evolving BTC/ETH bots beyond RSI rules, quant devs testing darwinian selection on volatile pairs, or AI builders prototyping self-evolving ai agent github workflows. Suits self-evolving agents openai fans wanting a cookbook for adversarial trading sims.

Verdict

Worth forking for darwinian trading experiments—solid CLI/dashboard and 111 tests mitigate its 46 stars and 1.0% credibility. Early beta, but quick evolutions (30s demos) make it a low-risk self-evolving agent github starter.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.