Looter Trading Journal

https://github.com/Seanamccormick7/DEHacksJournal

Inspiration

We have been trading stocks since we were sophomores in high school, learning to navigate the markets manually and learning on our own. This app draws inspiration from our yearning for personalized feedback without the expenses of a coach, and Fintellect AI.

What it does

Looter ingests your broker CSV and notes, runs them through OpenAI, and returns personalized, pattern-tracking trade critiques in seconds.

How we built it

A Next.js frontend pipes uploads to Supabase Postgres, triggers edge functions in TypeScript that call OpenAI, and streams the AI summary back via real-time channels.

Challenges we ran into

Normalizing inconsistent CSV formats, wrangling Supabase RLS, and keeping OpenAI API use low under hackathon-level load pushed our pacing and debugging skills.

Accomplishments that we're proud of

We delivered secure CSV import, live AI feedback, and an insight dashboard—all fully deployed.

What we learned

Tight prompt engineering, Supabase edge functions, and disciplined MVP scoping are critical when every minute counts.

What's next for Looter Trading Journal

We’ll integrate real-time alerts for bond-yield shifts, Powell’s Fed remarks, headline movers, and big-tech earnings while auto-tagging consistent high-growth, value stocks like Palantir.

Built With

• chatgpt
• cursor
• nextjs