# Changelog of Trustpilot Reviews Scraper — Ratings & Company Replies (`apivault_labs/trustpilot-reviews-scraper`) Actor

- **URL**: https://apify.com/apivault_labs/trustpilot-reviews-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/apivault_labs/trustpilot-reviews-scraper.md

## Changelog

### 2.1 — candidate, 10 October 2026

- Python collection contract with real review IDs, strict input validation, rating/language/country/date/verification/reply filters and field switches.
- Sequential pagination continues after filter misses, preserves source order, detects repeated pages, and records explicit caps and source errors.
- Review-only paid dataset; company profiles, diagnostics, counts and delivered timestamp checkpoint are free key-value records.
- Ordinary public HTTPS with bounded responses and redirects; explicit access denials stop collection.
- Honest availability, coverage, pricing, manual repeat-run and MCP documentation.
- Candidate is not a verified production release: source access is currently restricted.

### 2.0 — historical Go beta, not deployed as production

An experimental native Go implementation existed locally. Its advertised filters, cookie transport, company rows and full-history coverage were not verified in the existing production Actor. These files are excluded from candidate 2.1 deployment.

### 1.0 — existing production

- Multi-company, multi-page extraction, selected fields and content-based deduplication.
- Review event price $0.0005 since 31 May 2026.
- Baseline production build retained for rollback: 1.0.10.

### 2.2 — 2026-10-21

- Bound HTTP requests, redirects, streamed body reads and retries by the remaining page/run deadline.
- Respect earlier platform expiry with a reserve for diagnostics and final writes.
- Retain the reviewed candidate contract: useful-review billing and explicit FAILED/PARTIAL with free structured diagnostics. Source denials stop without alternate identity/session fallback.

### 2.4 — 2026-10-21 — candidate

- Reuse the sibling native reader’s Googlebot request header for public structured pages after a positive20-review probe; no authenticated API or cookies.
- Preserve strict review validation, deadline enforcement and explicit refusal diagnostics. Live exact-build QA required before promotion.