Breaking

Crypto Auto-Trading Bot PowerTrader_AI Released as Open Source

July 3, 2026 at 22:20 EDT

Open Source
Finance
AI Agents

An individual developer has released PowerTrader_AI, a crypto auto-trading bot reflecting nearly a decade of trading experience and AI design, as fully free open source under the Apache 2.0 license on GitHub. It combines a custom price-prediction AI with tiered Dollar-Cost Averaging (DCA) to run emotion-free automated trades.

Open Source · Apache 2.0

PowerTrader_AI: a fully autonomous crypto trading bot, released free

A bot that automates spot crypto trading using a custom price-prediction AI and tiered DCA — powered not by an LLM, but by an instance-based (kNN/kernel-style) predictor that learns online with no fixed pre-trained model.

1h–1wk

Timeframes of pattern history stored per asset

2×/24h

Maximum DCA triggers within 24 hours

Stop-loss orders — by design, none are used

Take-profit target, drawn to scale

The profit target is cut in half once dollar cost averaging kicks in.

Without DCA

2.5%

With DCA

A 0.5% trailing gap then follows any further rise.

How the prediction loop works

Store patterns

Full history across 1h–1wk timeframes

→

Forecast candle

Weighted average of nearest past patterns

→

Reweight online

Adjust weights against actual outcomes

Seen as a step forward

Praised as an example of autonomous AI trading — an experimental approach distinct from indicator-driven bots like Coinrule, WunderTrading and Freqtrade. Cross-platform: Windows, macOS, Linux and Raspberry Pi.

Handle with caution

No public backtest figures, win rate or ROI. Warnings over financial risk, sudden market swings and flawed reasoning. "Experimental and educational" — test small before any live use.

"You are fully responsible for all of your money."

The developer repeatedly cautions in the README that the tool is experimental, with no large-scale operational cases available — recommending small-scale testing first.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.