AutomationIn use

derivKAI

A rules-based trading bot with risk limits that cannot be overridden mid-session.

The problem

Synthetic index markets run around the clock. Trading them by hand leads to fatigue, hesitation and inconsistent position sizing.

Approach

A Python service connects to MetaTrader 5, applies volatility-adjusted position sizing and stop placement, and logs every closed trade to a Notion database for review. Monitoring uses read-only credentials, separate from the process that can place orders.

[ MT5 terminal / market feed ]
        │
[ Strategy engine ]
   ├── Risk and sizing filter ──> [ Order execution ]
   └── Notion trade journal

Key decisions

Read-only (investor) credentials for monitoring, kept separate from the execution process.
A hard daily loss limit is enforced in code before any new order is placed.
Automatic trade journaling removes manual bookkeeping and makes post-session review routine.

Highlights

Position sizing and stops are deterministic and rule-based.
Every trade is journaled automatically.
Personal project; not a product and not financial advice.

Stack

Python
MetaTrader 5 API
NumPy
WebSockets
Notion API