Building query string extractor for top 10k websites. Used model to find search box within home page and figuring out the search query string. It works for 80% of the websites. hope to release the API to beta users next week :) #buildinpublic #sideproject
searchresult.dev

I'm building my weekend side project, its search result API.
- completed AI based *web data record* extraction from search result pages.
- hosted my function at aws lambda.
- landing page is at searchresult.dev
#buildinpublic

This week end, I am planning to add more features (for watchlist.run) #MVP
- Signup/login
- Display Feed (new items in the given url)
I am taking a short break from twitter, to write some code, will do what I am good at #buildinpublic

#buildinpublic of my first product - watchlist.run
Day 2 of #100DaysOfCode
- None of the classic algorithms (MDR, DEPTA) works well on modern web pages, even after some fine tuning.
- Trying with visual similarity and the Gestalt laws of grouping. Hope this works.

#buildinpublic of my first product.
Day 1 of #100DaysOfCode

- Made a landing page: watchlist.run
- Selected some classic algorithms such as MDR, DEPTA and NET for list extraction - AI/Back end
- Will finalize the best algorithm in 3 to 4 days