Personal project

AI Trading Research System

A research assistant that studies the market each morning and texts me a short list of trade ideas.

AI automation Market research Built solo

What it is

Every trading day, this system reviews a set of stocks I follow, finds the few that match a strategy I tested carefully, and texts the ideas to my phone. In the background it keeps its market data current, re-checks the strategy against history each week, and tracks how past ideas turned out.

24stocks watched
Yearsof data tested
A few $per month to run
Dailyideas to my phone

How it works

Scan the market Spot a few setups Size by conviction Text me the ideas I place the trade
The system suggests; I make the final call.

A capable, more expensive assistant makes the judgment calls, while cheaper helpers handle the repetitive chores on a schedule. Splitting the work that way keeps the whole thing running for just a few dollars a month.

The idea behind it

It only acts on a few specific patterns that held up in testing, and it leans in harder when a signal is strong and eases off when it's weak. A lot of the work was the opposite of adding features — throwing out clever ideas that looked good on paper but lost money once I tested them. The simplest version won.

A careful approach

The system only suggests — I place every trade myself. On the first handful of real trades it's done well, but that's a small sample, so I treat it as a promising start rather than a promise.

Built with AI-assisted development — the design and product decisions are mine.

← Back to all work