# Changelog of Home Depot Scraper (`khadinakbar/home-depot-scraper`) Actor

- **URL**: https://apify.com/khadinakbar/home-depot-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/khadinakbar/home-depot-scraper.md

### 1.1 — 2026-06-21

- Reliability: SerpApi-outage runs now end SUCCEEDED with a terminal WARNING + one diagnostic record instead of FAILED. Provider outage is an external condition, not an actor failure.
- Fail-fast: per-request SerpApi timeout cut 60s → 25s and retries reduced (search 3 → 2, product detail 5 → 2) so a dead upstream engine no longer TIMES-OUT the whole run.
- Added an adaptive time-budget guard (reads ACTOR\_TIMEOUT\_AT, bails at ~90% of the budget, flushes partial results, force-exits) to prevent platform TIMED-OUT.
- Actor.fail is now reserved only for genuine code crashes and a missing/invalid SERPAPI\_KEY.

### 1.0 — 2026-06-19

- Initial release. Scrapes Home Depot products via SerpApi home\_depot (search) + home\_depot\_product (detail) engines.
- Four input modes auto-detected: keyword searches, product URLs, category/search URLs, and 9-digit product IDs.
- Returns flat product records: product\_id, title, brand, model\_number, price, originalPrice, rating, reviews, pickup/delivery availability, thumbnails, link. Optional full detail (specs, description, images) and per-review records.
- ZIP-aware pricing via deliveryZip. HTTP-only, no browser, no proxy — SerpApi absorbs Home Depot's Akamai protection for ~99% reliability.
- Reliability: SerpApi calls are serialized with exponential-backoff retry on transient engine errors (the 'try again later' signal) so a single flaky call never fails the run.
- Lifecycle: soft-fail on bad input, honest-fail only on total SerpApi exhaustion, upfront cost-cap log, triple-guarded charge counter, safePushData schema-fallback, RUN\_SUMMARY written to key-value store.
- Pay-per-event: actor-start $0.00005, product $0.005, review $0.002.