Crawl a whole site and audit it: the homepage, the sitemap and every internal link (robots.txt respected), up to max_pages, then 170+ Semrush-style checks graded error / warning / notice, a Site Health score (0–100) and 9 theme scores (crawlability, HTTPS, international SEO, Core Web Vitals, performance, internal linking, markup, AI search, on-page).
Billing: 1¢ per 10 pages fetched (1,000 pages = $1.00, 10,000 = $10.00), billed as it crawls; the link, image and script checks (up to 5,000 per crawl), the site checks (HTTPS, www, TLS, soft 404, llms.txt, AI fetchers) and Core Web Vitals on up to 10 pages are included at no extra charge. It stops as "partial" if the balance runs out. Confirm max_pages and the most it can cost with the user first.
Timing: about 3–6 pages a second, then a short link-check and scoring stage; Core Web Vitals follow a few minutes after completion (cwv_state "pending" until then; their theme only, never Site Health). It runs ~40 s in this call: a small site comes back complete, a bigger one returns status "running" and keeps crawling in the background — read it later with get_latest_crawl (free). One page instead: run_technical_audit.
Inputs: site_id (list_sites), max_pages (1–10000, default the whole site), seed_url (optional, on the site). Optional settings (source, include/exclude, ignore_params, subdomains, robots_agent, device, speed, cwv_urls): see each input; leave them out to crawl the whole site.
Returns: { run_id, site_id, status: running|complete|partial|failed, stage, coverage, pages_crawled, pages_found, pages_failed, skipped_by_rules, charged_cents, billed_cents (same), max_cost_cents, settings?, rejected?, site_health, counts: { errors, warnings, notices }, themes, cwv_state, checks: [{ id, severity, title, n, of }] (failing, worst first, ≤30), issues_summary (counts), top_problem_pages (≤5: { url, http_status, checks }), note, next_step }. While running, checks are provisional page counts so far. For one check's URLs, evidence and fix call get_crawl_issue (free).