I’m building an AI options trading assistant and publishing the paper-trading results publicly

I’ve been building an AI-powered options trading assistant called Ohyolo.

The idea is not to create another “signals” product. The goal is a controlled automation system where the user sets risk limits, connects a brokerage, and the AI manages QQQ options trades inside those limits.

Current flow:

* AI analyzes QQQ market data
* Picks CALL/PUT, strike, expiration, profit target, and stop-loss
* Risk filters check spread, volume, open interest, contract limits, and active trades
* SnapTrade handles brokerage connection and order placement
* The system stages the trade:
* entry order
* confirm fill
* place stop-loss
* monitor profit target
* AI reviews open trades during the day
* close if conditions change
* Dashboard tracks live PNL and closed trade history

I’m testing with paper trading right now, not live customer money.

So far in testing:

* One trade closed for about **+$1,220**
* One trade closed for about **-$270**
* Net so far: about **+$950**

Obviously that is way too small of a sample size to claim anything. The goal right now is testing execution reliability, order handling, risk controls, and transparent reporting.

I also created a public AI performance page so people can follow the paper-trading results over time. It shows trades taken, PNL, win rate, open trades, account leaderboard, and recent closed trades.

What metrics would you want to see on a public performance page before taking a tool like this seriously?

Also, what would make this feel more transparent instead of just looking like another trading bot?