On-page SEO Audit.

One call, every check

Most SEO audit tools make you load the dashboard, queue a crawl, then page through tabs of findings. /v1/actors/audit-onpage is the headless version: one POST, one JSON response with errors, warnings, and info findings grouped by rule. Drop it into any pipeline — agency reporting, pre-publish CI checks, change-detection on a competitor's flagship page — without touching a UI.

curl -X POST https://api.crawlcrawl.com/v1/actors/audit-onpage \
  -H "Authorization: Bearer crk_..." \
  -d '{"url":"https://your-client.com/page"}'

# → 200
{
  "actor": "audit-onpage",
  "url": "https://your-client.com/page",
  "elapsed_ms": 52,
  "data": {
    "summary": { "errors": 1, "warnings": 3, "info": 3 },
    "findings": [
      { "severity": "error",   "rule": "h1.missing",
        "message": "no <h1> on page" },
      { "severity": "warning", "rule": "meta.description-too-long",
        "message": "meta description is 207 chars; Google truncates at ~160",
        "count": 207 },
      { "severity": "info",    "rule": "heading.hierarchy-skip",
        "message": "heading hierarchy skips a level (e.g. h1 → h3)",
        "count": 3 }
    ]
  }
}

What's checked

The rule set is a Rust port of the well-tested python-seo-analyzer ruleset, expanded with checks for AI-search-readiness signals that became important in 2026.

• Title — missing, too short, too long, duplicate across pages (when used inside a crawl).
• Meta description — missing, too short, too long, duplicated.
• Headings — missing h1, multiple h1s, hierarchy skip, heading text too long.
• Canonical — missing, self-loop, points outside the domain, conflicts with hreflang.
• Robots — meta-robots, X-Robots-Tag header, robots.txt directive conflict.
• Open Graph + Twitter — missing required tags, image dimensions absent, type mismatch.
• Schema.org — invalid JSON-LD syntax, required-field omissions per type.
• Hreflang — return-link missing, self-reference missing, region code invalid.
• Images — alt-text missing, decorative-but-named, oversized.
• Performance hints — render-blocking script in head, missing preconnect for known font/CDN hosts.

When to call it

Three patterns dominate usage on crawlcrawl.

The first is agency reporting. An audit-onpage call against every top-50 page on a client site produces a per-URL findings table that goes straight into a monthly deliverable. Combine with /v1/crawls for full-site sweeps and /v1/actors/sitemap-audit to verify the sitemap matches what's actually crawlable.

The second is pre-publish CI. A GitHub Action that calls audit-onpage against the preview-URL of a PR catches "you removed the h1 in the redesign" before merge. Findings with severity: error fail the check; warnings post as a PR comment.

The third is competitor monitoring. Schedule audit-onpage on a competitor's flagship pages and you'll spot the day they fix their canonical structure or stop emitting JSON-LD.

Pricing

One page-credit per call. The $42 Studio tier includes 50,000 page-credits a month — enough to audit a 5,000-page site every Sunday for a year. See full pricing →

Where it fits