Market Tool
Python service + approval workflow
Python screens Alpaca market data and writes proposals to SQLite. Pushover delivers alerts, Tailscale keeps the dashboard private, and optional OpenAI review adds context.
Python reads Alpaca price and volume data.
SQLite records the proposal; Pushover sends an alert.
I review and approve. Unapproved proposals expire.
Python · Alpaca API · Pushover alerts · Tailscale · SQLite · OpenAI API
How it works
A Python service screens the stocks and ETFs that Alpaca lists as tradable, ranks the most active ones, and looks at one-minute price and volume bars for a short list. When a candidate qualifies it writes a proposal with the symbol, size, limit price and reasons. Nothing is sent to the broker until I review the proposal and approve it, and each proposal expires if I don’t. An optional AI review, run through the OpenAI API, can hold back a candidate or explain the risks, but it cannot approve or submit anything.
What I used
Python for the whole application, with no extra packages. The Alpaca API for market data and the brokerage connection. SQLite to keep a history of every proposal and decision. Pushover to send alerts to my phone. Tailscale so the dashboard is reachable from my phone without being public on the internet.
Why I built it
I wanted a way to watch the stock market without acting on impulse: something that does the scanning, shows its evidence, and leaves the decision to me.
What still needs checking
The starting strategy is deliberately simple and has not been proven. This is a tool for reviewing ideas, not a claim that it makes money, and it is not public.