Skip to main content
User lane: do not run local retraining or optimization. Use onboarding, select profile 1-5, pull approved model revisions from R2, and run trading.
This guide assumes you have a broker account with MetaTrader 5 and the MT5 API server already running. If you need to set that up first, see MT5 API server.

​
Prerequisites

  • Python 3.10–3.13 (3.13 recommended)
  • A broker MT5 account (VT Markets Cent for testing, any MT5 broker for live)
  • Cloudflare R2 credentials (to pull pre-trained models)

​
Steps

1

Clone the repository

git clone https://github.com/mokatific/novosky
cd novosky
2

Create a virtual environment

python3.11 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt
For detailed platform-specific instructions, see Installation.
3

Configure environment variables

cp .env.example .env
Edit .env with your values:
API_URL=http://<YOUR_MT5_SERVER_IP>:6542
API_TOKEN=<your_api_token>
CF_R2_ACCESS_KEY_ID=<your_r2_key_id>
CF_R2_SECRET_ACCESS_KEY=<your_r2_secret>

# Optional — disabling just silences notifications
TELEGRAM_TOKEN=<bot_token>
TELEGRAM_CHAT_ID=<chat_id>
For a full reference of all config files, see Configuration.
4

Run onboarding and choose your risk profile (1-5)

python scripts/onboarding.py --balance 500 --profile 3
Use --profile 1 to --profile 5 (or profile names) based on your risk tolerance.When onboarding asks about first optimization, choose n to skip retraining.
5

Pull the approved model revision from R2

python ml/r2_hub.py --pull --revision vYYYYMMDD-p3
Use the revision announced by your developer/operator for your selected profile.If you do not have direct R2 credentials, ask your operator for the approved revision and pull it from your managed environment.
6

Verify model compatibility

python3 -c "
import json
mc = json.load(open('models/model_compat.json'))
ml = json.load(open('ml_config.json'))
assert mc['feature_count'] == len(ml['features'])
print('OK:', mc['feature_count'], 'features')
"
This checks that the pulled models and feature config are in sync.
7

Run a true OOS backtest

Confirm the models work correctly on your machine before going live:
python backtest_config.py \
  --balance 500 --no-swap --leverage 500 \
  --spread 16.95 --oos-only --no-chart
You should see WR around 57–85% depending on the OOS window. See Backtesting for details.
8

Dry-run (diagnostic-only, safe)

python trading.py --dry
This runs the full trading loop — fetching candles, computing features, generating signals — but keeps side effects disabled.In --dry mode:
  • No real orders are sent.
  • No close/SLTP modification requests are sent.
  • No Telegram messages are sent.
  • Telegram command bot is not started.
  • No Supabase writes are made.
  • No runtime CSV/state files are written.
9

Go live

When you’re satisfied with the dry-run output:
python trading.py
For production deployments with auto-restart on crash and reboot persistence, use PM2:
npm install -g pm2
bash scripts/pm2-start.sh          # start with API readiness check
bash scripts/pm2-start.sh --setup  # first-time: enable auto-start on reboot
See First run for PM2 management commands.

​
Next steps

Configuration reference

All config.json and ml_config.json keys explained.

Brokers

Cent vs standard accounts, spread values per broker.

Telegram notifications

Every notification the bot sends and when.

Monitoring

Log files, dashboards, and model health checks.