In this video, we’ll see if Gemini 2.5 Pro can create a working trading bot in Python using the Alpaca API and VectorBT library. We’ll give Gemini prompt to implement trend, momentum, volatility, and volume indicators, use paper trading with Alpaca’s environment for testing, add some risk management using stop-loss orders and position sizing, use automated trading with Python at 1-minute intervals, and more.

💡 Tools & Libraries Used:
• Alpaca API (alpaca-py library)
• VectorBT for strategy development and backtesting
• Python 3

You can download the Jupyter Notebook used in this tutorial to follow along and practice on your own from tutorial on my site: https://superai.pl/courses.html

—
2/3 thumbnails were created with ChatGPT

—
Chapters generated by Gemini 2.5 Pro
0: 00 – Intro: Can Gemini 2.5 Pro Code a Trading Bot?
0: 09 – The Trading Bot Prompt for Gemini
0: 21 – Gemini Creates the Initial Bot Code
0: 31 – Setting Up the Python Environment (Alpaca & Libraries)
0: 48 – Implementing Gemini’s Code & Strategy
1: 05 – Gemini’s Trading Strategy Explained (SMA, RSI, Vol)
1: 34 – Disclaimer: Not Financial Advice! (Paper Trading Only)
1: 58 – Running the Bot: First Error Encountered
2: 01 – Debugging the Code with Gemini’s Help
2: 14 – Issue: Alpaca Data Lag & Gemini’s Solution (CCXT)
2: 29 – Modifying the Bot to Use CCXT/Binance Data
3: 07 – Evaluating Gemini’s Strategy Choice
3: 21 – Gemini Recommends Backtesting
3: 28 – Backtesting Gemini’s Initial Strategy (Using VectorBT)
4: 03 – Backtest Result: FAIL! (-99% Return)
4: 09 – Asking Gemini for Better Trading Strategies
4: 24 – Gemini Proposes 5 New Strategy Variations
4: 45 – Setting Up & Backtesting the 5 New Strategies
4: 51 – Backtest Results Summary (Strat 1 is the Winner?)
5: 00 – Updating the Live Bot with the Best Strategy (Strat 1)
5: 12 – Running the Final Live Bot (15-Minute Strategy)
5: 15 – Live Bot Executes First BUY Trade
5: 35 – Live Bot Executes First SELL Trade (at a Loss)
6: 22 – Current Market & Final Bot Status
6: 34 – Summary: Gemini’s Coding vs Strategy Performance
7: 35 – Conclusion & What’s Next? (SuperAI.pl)