We rebuilt every review-site scraper from the ground up with a consistent pipeline, dedup, and full pagination (no more "we got the first 10 and stopped"). Highlights:
- Yelp — 256 reviews via Yelp's GraphQL API through Web Unlocker
- YouTube — 500 comments fetched in ~2.4 seconds
- Reddit — 54 comments via stealth browser on old.reddit.com (free, no API key)
- Realtor.com — 412 reviews via
__NEXT_DATA__ - Zillow — 10 reviews via Web Unlocker with country bypass
- Udemy — dedicated scraper with improved scrolling
- Plus fixes for HomeStars, Fiverr, ProductHunt, Booking.com, G2, and more
Parallel batches of 5 pages, skip flaky pages instead of failing the whole import, dedupe by content hash.
Learn more: Import overview