SEO Audit Tool and Public SaaS Pricing & Link Auditor

Use auditMode: "pricing" for source-backed price candidates, metadata and CTA links from public SaaS pages. The Console example audits two publisher-owned pages with one safe link probe per page. Omit auditMode, or select seo, to keep the existing SEO crawler.

{
  "auditMode": "pricing",
  "startUrls": [
    "https://atesensoftware.com/apps/web/attachment-bundler/",
    "https://attachment-bundler.atesensoftware.com/"
  ],
  "maxPricingPages": 2,
  "maxPricingLinks": 1,
  "pricingRunSeconds": 30,
  "pricingRequestSeconds": 5
}

Pricing mode emits one pricing_page row per attempted page, up to the page cap: HTTP status, redacted final URL, title, canonical, detected noindex directives, priceCandidates with literal currency evidence and verbatim whitespace-normalized surrounding text, allowlisted JSON-LD Offer/PriceSpecification fields, explicit CTA labels/hrefs, and bounded link-probe results. $ alone remains an unspecified dollar currency. Billing intervals remain null; the context retains actual "per month" or "30 days" wording. No billing period, checkout amount or price correctness is inferred.

Empty pages and JavaScript app shells produce unknown_no_price_in_static_html. A 403, 401, 429 or detected challenge is unknown_blocked, never a confirmed broken link. Only observed 404/410 link responses are broken. Unprobed links, robots refusals, deadlines and network errors remain explicit unknowns. No login, purchase, OAuth, form submission or browser action is performed.

Pricing input	Default	Bound
maxPricingPages	2	1–20 explicit pages; no automatic page discovery
maxPricingLinks	1	0–20 safe CTA HEAD probes per page
pricingRunSeconds	30	1–60 seconds shared wall-time budget
pricingRequestSeconds	5	1–10 seconds per DNS/request, further limited by remaining run time

Pricing mode fetches public HTTP(S) HTML on standard ports. Credentials, all query-bearing URLs and common auth/payment/action routes are skipped. Private, loopback, link-local, multicast, reserved and metadata IPv4 addresses are blocked. Each connection is pinned to a freshly validated public IPv4 address; redirects are revalidated and capped at three. IPv6-only hosts are unsupported. Robots rules apply at each destination, including link probes. Requests start at least one second apart per hostname; longer declared crawl delays are honored. Responses are capped at 1 MiB, without automatic retries, proxies or challenge workarounds. Compressed responses that ignore the identity-encoding request remain unknown. These safeguards describe pricing mode; legacy SEO mode retains its existing behavior.

Static extraction excludes scripts, templates, hidden attributes and inline-hidden elements. It does not compute CSS layout or execute JavaScript. Computed visibility, dynamic prices, country-specific prices, taxes and customer offers are outside its evidence. This is an experimental utility without a market-demand or revenue claim.

Pricing mode uses the existing result event for successfully audited HTML pages. Failed, blocked or non-HTML page rows have no custom result event. Pricing configuration is unchanged; the live Pricing tab is authoritative. RUN_SUMMARY records actual page/request counts and budget/time stops. The legacy CSV outputs below apply to SEO mode.

An explicit zero event-charge budget produces no billable page rows. Pricing mode also stops before fetching a page when that budget is zero. API and CLI callers should set maxTotalChargeUsd explicitly and check the actual run options; the event cap is distinct from any platform-usage charges shown for the run.

Existing SEO mode

Crawl any website and get a full technical SEO audit for every page plus a broken link checker in the same run, at $1.50 per 1,000 pages ($0.0015 per audited page; site summary, issue and status rows are free). You get a 0–100 SEO score with a letter grade, broken links and images (4xx/5xx and redirect chains), meta/Open Graph/Twitter checks, structured data (JSON-LD) validation, Core Web Vitals risk hints, robots.txt and sitemap checks, and ready-made CSV reports.

No browser, no proxy and no setup. Paste a list of domains and press Start.

What does SEO Audit Tool do?

For each website you enter, the actor:

1. Fetches the start page and checks site-wide signals: robots.txt, XML sitemap (URL count), HTTP→HTTPS redirect, www/non-www consistency, HSTS and soft-404 handling.
2. Crawls internal pages (breadth-first, follows robots.txt) up to your page limit.
3. Audits every HTML page against 55+ checks in 10 categories and scores it from 0 to 100 (A–F).
4. Checks every internal and external link and every image for 4xx/5xx errors and redirect chains.
5. Finds duplicate titles and meta descriptions across the site.
6. Writes one row per page, one summary row per site and three CSV files (pages, issues, broken links).

Why use this SEO audit tool?

	This actor	Typical browser-based SEO audit actor
Price per 1,000 pages	$1.50	$30–40
Broken link + image checks	✅ internal and external	often extra or missing
Structured data validation	✅ JSON-LD, microdata, RDFa, required properties	basic
Core Web Vitals risk hints	✅ LCP / CLS / INP risk per page	—
Score per category + grade	✅ 10 categories	single score
CSV reports	✅ PAGES.csv, ISSUES.csv, BROKEN_LINKS.csv	JSON only
Memory	512 MB	2–4 GB
Failed or blocked sites	free status row, not charged	often charged or the run fails

What does it check?

• Meta: title (missing, length, pixel width, duplicates), meta description (missing, length, duplicates), canonical (missing, points elsewhere, relative, multiple), meta robots and X-Robots-Tag (noindex/nofollow), charset, lang, favicon, hreflang.
• Headings & content: missing or multiple H1, skipped heading levels, word count, thin content, text-to-HTML ratio.
• Technical: HTTP status, redirect chains, HTTPS, mixed content, URL length/uppercase/underscores/parameters, indexability.
• Mobile: viewport meta tag.
• Performance hints: TTFB, HTML size, DOM size, render-blocking scripts and stylesheets, number of third-party domains, compression, document.write, missing preconnect.
• Images: missing alt text, missing width/height (CLS risk), lazy-loaded hero image (LCP risk), legacy formats, broken images.
• Links: broken internal/external links, redirected links, empty anchor text, nofollow internal links, few internal links.
• Structured data: JSON-LD parse errors, schema types, missing required properties (Product, Offer, Article, Organization, LocalBusiness, Event, Recipe, FAQPage, BreadcrumbList, JobPosting and more).
• Social: Open Graph (title, description, image) and Twitter Card tags.

Each issue has a severity (critical / warning / notice), a category, a message and a fix hint.

Core Web Vitals hints

The actor does not run a browser, so it cannot measure real LCP, CLS or INP. Instead it estimates the risk (low / medium / high) of each metric from the HTML: slow server response, render-blocking resources, lazy-loaded hero images, images without dimensions, heavy DOM, many third-party scripts and so on. Use the hints to find pages to fix. For measured values use PageSpeed Insights or the Chrome UX Report.

How much does it cost?

Pay per event. You pay only for audited HTML pages: $0.0015 per page ($1.50 per 1,000 pages), the same on every Apify plan, with no start fee.

These rows are free: site summary rows, issue rows, status rows for invalid, unreachable, blocked or non-HTML URLs, and pages that failed to load. Link and image checks are included in the page price.

Example: auditing 20 websites with 50 pages each (1,000 pages) costs about $1.50.

Input

Field	Default	Description
startUrls	–	Websites or pages to audit. Domains without https:// are fine.
maxPagesPerSite	25	Maximum pages audited (and charged) per website.
crawl	true	Follow internal links. Off = audit only the listed URLs.
maxDepth	10	Maximum link depth from the start page.
checkLinks / checkExternalLinks	true	Check links for 4xx/5xx errors.
maxLinkChecksPerSite	500	Unique links checked per site.
checkImages / maxImageChecksPerSite	true / 200	Check images for errors.
respectRobotsTxt	true	Skip URLs disallowed by robots.txt.
includeSubdomains	false	Treat subdomains as part of the site.
includeUrlPatterns / excludeUrlPatterns	–	Only crawl, or skip, URLs that contain these strings.
ignoreQueryParameters	false	Treat ?a=1 variants as the same page.
outputIssueRows	false	Also write one dataset row per issue.
maxConcurrency	4	Parallel requests per run.

Example input:

{
  "startUrls": ["https://example.com", "books.toscrape.com"],
  "maxPagesPerSite": 50,
  "checkExternalLinks": true
}

Output

The dataset contains rows of four types (use the views in the Output tab):

• page: one row per audited page, with URL, status, title, description, canonical, H1, word count, links, images, schema types, score, grade, per-category scores (scoreMeta, scoreTechnical, …), cwvLcpRisk / cwvClsRisk / cwvInpRisk, issueIds, issuesText, brokenLinks and the full issues array.
• site: one summary per website, with siteScore, grade, average category scores, top issues, broken link count, duplicate titles and descriptions, robots.txt, sitemap, HTTPS, www and soft-404 results.
• issue: one row per issue (only if outputIssueRows is on).
• status: free rows for invalid input, unreachable, blocked, HTTP error or non-HTML start URLs.

Example page row (shortened):

{
  "type": "page",
  "url": "https://books.toscrape.com/",
  "statusCode": 200,
  "score": 71,
  "grade": "C",
  "title": "All products | Books to Scrape - Sandbox",
  "metaDescription": null,
  "h1": "All products",
  "wordCount": 412,
  "cwvLcpRisk": "medium",
  "issueIds": "DESC_MISSING; CANONICAL_MISSING; IMG_NO_DIMENSIONS; OG_MISSING",
  "brokenLinkCount": 0
}

CSV reports

The key-value store of every run contains spreadsheet-ready files (UTF-8, open directly in Excel or Google Sheets):

• PAGES.csv: one line per page with all scores and the main fields.
• ISSUES.csv: one line per issue with page, severity, category, message and fix hint.
• BROKEN_LINKS.csv: source page, broken URL, status code and link type.
• RUN_SUMMARY: JSON totals for the run.

Tips

• Monitor your site weekly: create a task with your domains and add a schedule (for example 0 6 * * 1). Compare siteScore and brokenLinkCount between runs.
• Start with maxPagesPerSite: 25 to check a site cheaply, then raise the limit for a full crawl.
• Turn off checkExternalLinks for very large sites to finish faster.
• Some sites block automated requests. The actor then writes a free blocked status row and charges nothing.

FAQ

Does it use a browser? No. It downloads raw HTML with a browser-like HTTP client, so it is fast and cheap. Content rendered only by JavaScript is not seen, and neither is it by many crawlers.

Is it legal? The actor reads only public pages and follows robots.txt by default. Audit sites you own or have permission to analyse.

Something is missing or wrong? Open an issue on the Issues tab and we will look at it.

Pricing

Pay per event: $0.0015 per result ($1.50 per 1,000 audited pages). Each audited page row is one result event. Site summary rows, issue rows, status rows, error rows and empty lookups are not charged. The run stops when it reaches the maximum cost per run. No Apify proxy and no login.