Pre-publication review pass.
- Removed
category and reviewCount from the output. Blinkit returns an empty category hierarchy in both search and category responses, and its web API exposes no review count, so both were permanently null. subcategory (Blinkit's product type) and sourceTerm (which search or category a row came from) cover the same ground with real values.
- Pricing set to $0.0015 per product scraped ($1.50 per 1,000), plus the existing $0.00005 run-start charge.
- Rewrote the README around what the data is worth, with verified multi-location and sponsored-placement measurements.
- Regrouped the input form: search terms, locations and the result limit come first and are enough on their own; categories, product URLs, filters and advanced options follow.
- Added a fourth dataset view, Organic results only, for separating paid placements from organic ranking. The default view now leads with inventory, rating and delivery ETA alongside pricing.
First release.
- Search Blinkit by keyword, browse categories by name or URL, and fetch products by product URL
- Multi-location scraping with city name, PIN code or coordinate input
- Numeric pricing with MRP, discount amount/percent and per-unit price normalisation
- Live per-store inventory counts, stock status and delivery ETA
- Ratings and sponsored-placement (
isSponsored) detection
- Serviceability validation per location, with unserviceable locations skipped rather than failing the run
- Bounded retries with exponential backoff and residential proxy rotation; a failure on one query or category does not end the run
- Pure HTTP collection — no browser — running comfortably in 256 MB
