# Changelog of 🍿 Netflix Search Scraper (`data_minds/netflix-search-scraper`) Actor

- **URL**: https://apify.com/data\_minds/netflix-search-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/data\_minds/netflix-search-scraper.md

## Changelog

### v1.0 — 2026-09-01

First release.

- Netflix catalogue search by plain keyword, by Netflix search URL, or by a link to one specific title — all accepted together in one bulk input.
- Full catalogue filters: movies/series, release-year range, IMDb rating range, 37 regions with four region-match modes, audio language, subtitle language, sort order, and "leaving soon" / "added in the last 24 hours".
- Region-by-region streaming availability for every title, with added date, expiry date, audio and subtitle languages and season coverage — not the truncated first-ten list the search page returns.
- IMDb layer per title: rating, votes, metascore, genres, awards, full plot, IMDb link and Top-250 flags.
- Opt-in cast & crew and full season/episode breakdowns.
- Automatic network fallback: direct → datacenter proxy → residential proxy (3 retries, sticky), with a residential safety net kept underneath a user-pinned datacenter route.
- Automatic source fallback: if the main catalogue index stops answering, the run switches to a backup index and still returns data instead of failing, with every degraded record labelled in its own `enrichmentErrors` column.
- Nine live dataset views: Overview, Ratings & IMDb, Where it streams, Cast & genres, Seasons & episodes, Artwork & story, Links & ids, Source & quality, Errors.
- Live dataset writes, graceful abort/migration handling, and a run summary in the key-value store.
- Drop-in compatible with the input field names other Netflix Actors on the Store use, including their snake\_case filter names.
- Title look-ups match strictly: a link naming one film never comes back as a different film. If no candidate actually matches the name, the input is reported as unmatched rather than resolved to the search engine's first row.
- The catalogue filters are enforced on results from the backup index too, which only accepts a search term and a region — so "movies only, 2015 onward, rated 7+" holds whichever index served the run.
- Requests are paced by the gap between request starts, not only capped by concurrency, since launch rate is what a catalogue actually measures.
- The source-outage breaker is tied to the per-request retry budget rather than a larger fixed constant, so a single-term run reaches the backup index instead of returning an empty result set.
- Dataset field names and the `source` value are neutral, self-describing names rather than the upstream indexes' internal naming.