~/meetings / 2026-03-19
Meeting #9

> 2026-03-19

First meeting after Whit's wedding and honeymoon. The group kicked off the Store Assistant — Product Location Lookup project: designing the full workflow from Siri shortcut to web scraping HEB.com and returning an aisle location. Whit scaffolded the new unified server and attempted to get Puppeteer running, hitting setup issues most of the session.

3 attendees(Whit, Lawrence, Nate)
##topics covered
  1. 01.Store Assistant concept — locate any HEB product by voice via Siri, same idea as the seafood SKU shortcut
  2. 02.Full workflow design: Siri captures GPS location + product name → sends to server → server scrapes HEB.com → returns aisle number
  3. 03.Geolocation challenge — server lives in Virginia, not the store; must use the phone's GPS to select the correct store on HEB's website before searching
  4. 04.Web scraping explained — using a server to navigate a website, click buttons, and extract data
  5. 05.Puppeteer introduced — headless Chrome API for browser automation; page.click(), page.goto() etc.
  6. 06.HEB doesn't have a public API — web scraping is the workaround
  7. 07.API vs web scraping distinction — APIs are like a waiter, web scraping is going into the kitchen yourself
  8. 08.Stock availability as a bonus feature — the same page element that shows aisle also shows 'out of stock'
  9. 09.Unified server plan — combine seafood-sku-lookup and product location into one 'Store System' server with two routes
  10. 10.Two API routes: /api/sku/:fishName (existing) and /api/location (new product lookup)
  11. 11.async/await explained — JavaScript needs to wait for web responses before processing; async prevents the code from running ahead
  12. 12.URL parameters vs request body for passing data to the server
  13. 13.Boilerplate code explained — the baseline setup code every project starts with
  14. 14.Code comments — // syntax, grayed out, ignored at runtime
  15. 15.Puppeteer install and Chrome setup struggles — Linux compatibility issues, 'Could not find Chrome' errors
  16. 16.Store map/SVG question — whether the server could return the visual in-store map (uncertain; it's an SVG, not a plain image)
  17. 17.Chat HEB — HEB's internal AI tool pilot mentioned; context around why the company restricts outside AI use
  18. 18.AI data privacy — risk of sharing store sales data with ChatGPT (competitor access, location data)
  19. 19.Omarchy name origin — Japanese for 'chef's kiss' combined with Arch Linux as a pun by DHH
##decisions & next steps
  • →Combine seafood-sku-lookup into a new unified 'Store System' server
  • →Start with plain text aisle output; store map image is a stretch goal
  • →Whit to prep Puppeteer setup and test basic scraping before next meeting
← meeting #8meeting #10 →