# Changelog of Semrush Full API SEO Scraper - Traffic, Authority, Backlinks (`pro100chok/semrush-scraper`) Actor

- **URL**: https://apify.com/pro100chok/semrush-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/pro100chok/semrush-scraper.md

## Changelog

### 0.2.13 (2026-09-22)

- Fixed runs failing at startup with `ConnectionError: Access to requested hostname is forbidden` on accounts without access to Apify Residential proxy — even when Moz was turned off. The Moz proxy is now set up only when `include_moz` is on, paying users get Moz through the built-in premium proxy, and if Apify Proxy is unavailable the run carries on (Moz columns stay empty) instead of crashing.

### 0.2.12 (2026-08-23)

- Run logs are now fully in English — a handful of session, retry, and account-pool messages were still printed in Russian.

### 0.2.11 (2026-08-22)

- Fixed accounts with a dead session staying in rotation: Semrush answers a banned account exactly as it answers an anonymous visitor (HTTP 200 with no user data, or `Operation forbidden` on keyword requests), and the Actor treated that as a temporary glitch instead of a dead session. Such accounts are now retired immediately, so runs stop wasting attempts on them and switch to a working account sooner.

### 0.2.9 (2026-08-17)

- Fixed domains silently missing from the output: when a proxy IP went bad (`SSL_connect: Connection closed abruptly`), all five retries reused that same IP and the domain was dropped without a dataset record. Retries now fall back to a different sticky IP, and a domain that still fails goes back into the queue for another worker instead of being discarded.

### 0.2.8 (2026-08-17)

- Fixed empty `moz_domain_authority` / `moz_spam_score`: moz.com started blocking every Chrome TLS fingerprint, so most Moz requests came back 403. The scraper now impersonates Safari/Firefox profiles, which pass reliably.
- Raised request timeouts from 15 s to 30 s — slow residential IPs were failing whole domains on `curl (28) Operation timed out`.
- Made it explicit when Moz is disabled: the run log now says the Moz columns will stay empty, and the input option spells out which fields it fills.
- Removed Moz `top_competitors`, `top_ranking_keywords` and `ranking_distribution` from the output and the docs — moz.com no longer publishes those sections on its free domain-analysis page.

### 0.2 (2026-07-07)

- Improved network reliability and consistency of requests, especially on larger runs.

### 0.1 (2026-07-07)

- First release of the Semrush All-in-One Scraper with four modes selectable via the `mode` input: Domain Overview, Keyword Research, SEO Audit, and Top Websites.
- Domain Overview: Semrush data per domain, including authority score, total/follow/nofollow backlinks, referring domains, organic and paid traffic, keywords and traffic cost, AI search visibility (ChatGPT / AI Overview / AI Mode / Gemini), organic and paid competitors, SERP feature distribution, and optional Moz domain analysis (DA, spam score, top pages and linking domains).
- Keyword Research: search volume, CPC, competition, keyword difficulty, decoded search intents, related keywords, question keywords, topic clusters, ad copies, live SERP, 12-month volume trend, and volume by country.
- SEO Audit: on-page checks (title, meta description, robots, canonical, headings, word count, internal/external links, images, structured data, robots.txt and sitemap presence) enriched with Semrush organic context.
- Top Websites: top 100 trending websites with total visits, month-over-month and year-over-year change, and desktop/mobile split, filterable by country and industry.
- Dataset schema with per-mode Output views (Domain Overview, Keyword Research, SEO Audit, Top Websites) and export to JSON, CSV, and Excel.