Built after a real UK-wide Google Places extraction job. Focuses on what actually breaks at scale: pagination, deduplication, rate limits, long runs, resumability. Official API only. No scraping.