Installation

🦉 AthenaAI is an AI-powered ensemble machine-learning trading bot for PocketOption, built on BinaryOptionsToolsV2 → Get it on GitLab
💬 Need help? Join the Chipa Discord · ⚡ Go no-code with ChipaEditor

AthenaAI is a Python trading bot for PocketOption. Installation is three steps: Python dependencies from requirements.txt, a manual install of the BinaryOptionsToolsV2 broker client from its release wheel, and (optionally) PyTorch if you want GPU-accelerated neural-network training.

Prerequisites

• Python 3.9 or newer (the codebase uses modern type hints such as deque[Candle]).
• pip available in the same environment.
• A PocketOption account (needed later for live trading, not for offline training).

A virtual environment is recommended:

python -m venv .venv
source .venv/bin/activate        # Linux/macOS

python -m venv .venv
.\.venv\Scripts\Activate.ps1     # Windows PowerShell

1. Install Python dependencies

From the repository root:

pip install -r requirements.txt

This installs the two required libraries:

If scikit-learn is missing at runtime, the bot prints WARNING: scikit-learn not found. Install with: pip install scikit-learn and the ensemble falls back to a non-learning stub — so treat this dependency as mandatory.

2. Install BinaryOptionsToolsV2 (manual)

BinaryOptionsToolsV2 is the PocketOption client library. It is not on PyPI — download the wheel for your platform and Python version from its releases page:

https://gitlab.chipatrade.com/chipadevorg/BinaryOptionsTools-v2/-/releases

Then install it:

pip install path/to/BinaryOptionsToolsV2-<version>-<platform>.whl

Without it, python main.py exits immediately with:

ERROR: BinaryOptionsToolsV2 not installed.

Note: the standalone trainer (python train.py) does not need this package — you can train a brain on a machine that never connects to PocketOption. See Training.

3. Optional: PyTorch for GPU training

PyTorch enables the neural (MLP) model in the trainer. It is optional — everything else runs on CPU with scikit-learn.

• CPU-only: pip install torch

• NVIDIA GPU (CUDA build): follow https://pytorch.org/get-started/locally/, for example:

  pip install torch --index-url https://download.pytorch.org/whl/cu128

• Apple Silicon: plain pip install torch — the MPS backend is detected automatically.

See GPU training for device detection details and portability notes.

Per-OS notes

Windows

• Use PowerShell; set environment variables with $env:PO_SSID='...' (or use a .env file — see Quickstart).
• Pick the win_amd64 wheel of BinaryOptionsToolsV2 matching your Python minor version (e.g. cp311).
• Sample generation during training uses multiple processes; run scripts from a normal entry point (both train.py and main.py already guard with if __name__ == "__main__").

Linux

• Install python3-venv/python3-pip from your distribution if missing.
• Pick the manylinux wheel matching your Python version.

macOS

• On Apple Silicon, use an arm64 Python and the corresponding wheel.
• pip install torch gives you Metal (MPS) GPU acceleration for the neural model with no extra steps.

Verifying the install

Check each layer:

python -c "import numpy, sklearn; print('core deps OK')"
python -c "from BinaryOptionsToolsV2.pocketoption import PocketOptionAsync; print('broker client OK')"
python train.py --check-gpu

--check-gpu prints the detected training device without training anything, e.g.:

Device: cuda  (GPU: NVIDIA GeForce RTX 4070 (12.0 GB VRAM))

or, without PyTorch:

Device: cpu  (CPU (PyTorch not installed — pip install torch for GPU training))

Finally, running python main.py without configuration should fail cleanly with the PO_SSID is not set message — that confirms the entry point itself works.

Next steps

• Quickstart — configure your session ID and run the bot end to end.
• Configuration — the three configuration layers.
• Troubleshooting — if any step above failed.

💬 Stuck during setup? The Chipa Discord is the fastest place to get help — and if you’d rather skip the code entirely, build your strategy visually with ChipaEditor.