# Changelog of SEO Audit & Sitemap Checker - 404s, Redirects, Meta Tags (`syntrovia/sitemap-seo-check-http`) Actor

- **URL**: https://apify.com/syntrovia/sitemap-seo-check-http/changelog.md
- **Full Actor documentation**: https://apify.com/syntrovia/sitemap-seo-check-http.md

## Changelog

### 0.1 (2026-11-15, not yet published)

- Build 0.1.5 (QA round 2): a site behind a login wall no longer makes unbounded requests. Skipped login pages and rows that fail the output check count toward Max pages, and the run stops with `login-wall` (SUCCEEDED with what was saved, nothing charged for skipped pages) after 5 login pages in a row or when more than half of the first 20 checked pages are login pages. README responsible-use paragraph worded to match: individual login pages are skipped, login walls stop the run.

- Build 0.1.4 (QA and legal round 1): a login popup, sidebar login box or CAPTCHA on a contact or newsletter form no longer stops the audit; such pages are saved with the new `has-login-form` or `has-captcha` issue. Only a CAPTCHA or bot-check interstitial or HTTP 401/403/407/429 stops the run; a login page reached by a link is skipped (not saved, not charged) and the crawl goes on. Sign-up, register, join and similar account URLs are not requested. www. and the bare domain share one speed limit. Bounded memory for seen URLs (hashes) and the queue, URLs over 2,048 characters skipped, checkpoint records split under 1 MB, redirect targets kept after a migration, large JSON-LD no longer reported invalid, every row checked against the output schema before it is saved. README wording from the legal review.

- Build 0.1.3: fixed a count that could stop an audit a few pages before Max pages, or before the budget was used up, while a batch was being saved.

- First version. Audit mode: reads robots.txt and the sitemaps, then checks the start URL, the sitemap URLs and same-site links (up to 5,000 pages), one row per requested URL with status, redirect target, title, meta description, H1, canonical, robots directives, indexability, language, hreflang, viewport, Open Graph, JSON-LD types, word count, link and image counts, and 35+ issue codes.

- Sitemap mode: lists every URL of the sitemaps (up to 100,000) with lastmod, changefreq, priority and robots.txt status, without requesting any page.

- Sitemaps: indexes up to 3 levels, gzip files detected by their content, plain-text sitemaps, protocol size limits, up to 50 files per run.

- robots.txt per RFC 9309 on every run (own group or `*`, longest match, 404 = allow, 5xx = disallow); Crawl-delay honoured (one request at a time, at least the delay apart; up to 10 s in audit mode, 60 s in sitemap mode).

- Politeness and safety: at most 2 requests at a time per host, 10-120 requests per minute (default 60), honest User-Agent, no proxies, no browser; stops on 401/403/407/429, CAPTCHA, bot-check or login pages (changed in 0.1.4/0.1.5: login pages are skipped, login walls stop the run); refuses local, private, link-local and reserved addresses, also after redirects and at DNS connect time.

- Privacy: e-mail addresses and phone numbers (also international formats) masked in title, meta description and H1; no body text, link texts, mailto: or tel: links; URLs carrying contact details are not requested.

- Pay-per-event with exactly three events: `run-start` (with the first saved result), `page` and `sitemap-url`. The budget is checked before every request and every charge; a run that saves nothing costs nothing.

- Stops cleanly 2 minutes before the run timeout (`time-limit`), so long audits end as SUCCEEDED with what they saved.

- Results saved in batches of up to 25 pages or every 15 seconds; migration and abort stop at a batch boundary, so no page is saved or charged twice.