Inspiration

Being a retail investor, I always find it challenging to marry several stock performance signals, stock news sentiments and come up with a recommendation on a stock. And even more challenging is to scale it to hundreds of stocks and cherry pick what I want to invest in.

What it does

Recostock helps you discover top U.S. stocks by sector using smart financial analysis, earnings trends, and real-time news sentiment for data-driven investment decisions.

How we built it

Idea/requirement driven approach:We communicated directly with Bolt by submitting detailed prompts, describing desired functionality, UX views, and improvement requests. Interactive development: Bolt was easy to use and made the development cycle interactive. Can see code and preview right away. Reinforced learning: Bolt took the suggestions and feedback well and made precise adjustments to the code. It didn't made any unintended changes and disturbed existing functionality! Validation: Finally as we started to check for the data issues and suggested changes to bolt, it was quick in pointing existing API cannot serice the purpose and suggested an additional API which we leveraged and got the data fixed!!!

Challenges we ran into

  1. Certain times the code hallucinated and confidently gave wrong data answers. Ex: It took daily high/low data for 52wk high or low. I have to suggest bolt to fix it and got right answers.
  2. Token limits on Bolt. Thankfully helped it to ease over weekend with increased limits!

Accomplishments that we're proud of

  1. Built a functional app in just few hours against several weeks which it otherwise takes.
  2. Had deployable code and complete working model right after providing prompts!!!!
  3. Doesn't have to learn each tech stack but just focus on functionality and idea.
  4. Vibe coding at its best!

What we learned

  1. Learnt vibe coding, good prompting techniques like few shot prompting, etc.
  2. How to use bolt and how simple the programming is now!
  3. That we don't need to master programming language. Just focus on idea, functionality and propting.

What's next for RecoStock

Just want it to be a personal stock recommender for any retail investor!!!!. Update the code with more functional aspects and signals and make it a real product to use.

Built With

  • alphavantageapi
  • bolt.new
  • css
  • lucide
  • netlify
  • polygon.ioapi
  • react
  • tailwind
  • typescript
  • vite
  • yahoofinanceapi
Share this project:

Updates