# Changelog of TikTok Shop Shoppable Video & Creative Finder (`nexascout/tiktok-shop-shoppable-video-creative-finder`) Actor

- **URL**: https://apify.com/nexascout/tiktok-shop-shoppable-video-creative-finder/changelog.md
- **Full Actor documentation**: https://apify.com/nexascout/tiktok-shop-shoppable-video-creative-finder.md

## Changelog

### 0.1.3 — September 15, 2026

- Stop queued video detail/search navigations once confirmed delivered rows reach
  the per-product limit, using RequestQueue.fetchNextRequest and skipNavigation.
- Keep other products active and drain skipped detail counters cleanly.
- Avoid new video discovery scheduling for products already at their limit.
- Add RESULT\_LIMIT\_REACHED and skippedVideoRequestsAtLimit diagnostics.
- Preserve extraction, relevance, scoring, fields, view thresholds and fallback
  behavior before the output limit; existing in-flight pages may finish.
- 62 offline tests pass. No paid live test or Store pricing change in this release.
- Add a pricing sensitivity review using the actual $0.435 / 5-result owner test.

### 0.1.1

- Replaces the default GEENEU example with the live-verified HPBS product.
- The HPBS owner test returned six public creative rows with seven successful requests and no failures.
- Keeps extraction, scoring, schemas, and evidence boundaries unchanged.

### 0.1.0

- Initial TikTok Shop Shoppable Video & Creative Finder release.
- Reuses the verified TikTok Shop product and public-video extraction pipeline.
- Returns one dataset row per public video/product association.
- Adds deterministic caption hook, CTA, format, angle, and hashtag analysis.
- Adds transparent creative scoring and association confidence.
- Adds hard browser-page budget and Limited-permission Actor configuration.
- Adds input, dataset, and output schemas plus Store/MCP documentation.

## 0.1.2 — September 14, 2026

- Save structured creative rows during extraction, serialize duplicate writes, and
  restore already delivered associations from the dataset on restart.
- Keep the existing automatic default-dataset event; do not double-charge results.
- Provide a separate sorted `RANKED_CREATIVES` JSON export; default dataset streams
  in discovery order. Per-query slots go to first products with verified creatives.
- Block images and binary playback downloads; warm up the homepage only on retry;
  reduce request retries from two to one.
- Add short title-based searches independent of seller names, a 30-second default
  indexed-search time allowance, and at most two additional snippet probes.
- Fall back to browser search if indexed detail pages produce no delivered results.
- Publish qualifying structured search-page videos without redundant detail fetches.
- Keep snippets out of paid rows and recheck relevance using actual video content.
- Emit compact step diagnostics and explicit partial/empty/blocked outcomes.
- Align runtime limits with the schema. Document scheduled-page vs dollar limits.
- Validation: 57 offline tests, including three full pipeline scenarios; JS syntax;
  input, dataset and output schemas using official Apify validators with networking
  disabled. CLI validation was stopped by automatic review due to Segment telemetry;
  its standalone schema libraries were used instead, without running the CLI.
- No paid Actor runs or deployment performed. Live recovery rate, traffic savings,
  and compatibility with current site responses are not yet verified.