On-Page SEO Auditor — Meta Tags, OG, Schema & Score

Bulk on-page SEO audit API for agencies and migrations. Paste URLs and get a 0–100 seoScore with prioritized issues: titles, meta descriptions, canonicals, robots, Open Graph, Twitter cards, headings, JSON-LD schema, image alt coverage and word count — about $5 per 1,000 pages.

Why users pick this Actor

• Clear pay-per-event pricing — you only pay for successful results (~$0.005/page)
• Structured JSON ready for CRM, sheets, Zapier/Make/n8n
• Works with AI agents via Apify MCP
• Honest limitations documented — no fake data

SEO Page Auditor is a lightweight, browser-free Actor built for bulk checks. Paste your URLs, run it, and get one structured audit per page that you can sort by score, filter by issue code, or feed into a dashboard, spreadsheet or AI agent.

What does it check?

• Title presence and length (30–60 characters recommended)
• Meta description presence and length (70–160)
• Canonical URL, including cross-domain canonicals
• Robots directives from <meta name="robots">, googlebot and the X-Robots-Tag header (noindex / nofollow)
• Open Graph and Twitter Card tags
• Headings h1–h3 (missing or multiple H1s)
• Structured data: JSON-LD @type values and microdata types
• Images: total, with alt text, decorative (empty alt), missing alt, and alt coverage %
• Word count of the main content (nav, header and footer excluded when possible)
• Technical basics: viewport tag, lang attribute, HTTPS, HTTP status, internal/external link counts

Why use this Actor?

	Manual checks / browser extensions	Enterprise SEO suites	SEO Page Auditor
Bulk URL lists	Painful	Yes, but account and crawl limits	Paste hundreds of URLs
Machine-readable output	Copy-paste	Export quotas	JSON/CSV/Excel + API
Pricing	Free but slow	Monthly subscription	Pay per page, no subscription
Automation / scheduling	No	Limited	Schedules, webhooks, API, MCP
Consistent scoring	Varies	Proprietary	Transparent penalty model

Input

Field	Type	Description
urls	array (required)	Pages to audit. Each URL is audited individually — the Actor does not crawl links.
proxyConfiguration	object	Off by default. Enable only if a site blocks your IP. No captcha needed.

{
    "urls": [{ "url": "https://apify.com" }, { "url": "https://crawlee.dev" }],
    "proxyConfiguration": { "useApifyProxy": false }
}

Output

{
    "url": "https://crawlee.dev/",
    "finalUrl": null,
    "statusCode": 200,
    "title": "Crawlee · Build reliable crawlers. Fast.",
    "titleLength": 39,
    "metaDescription": "Crawlee covers your crawling and scraping end-to-end and helps you build reliable scrapers. Fast.",
    "canonical": "https://crawlee.dev/",
    "robots": null,
    "lang": "en",
    "og": { "title": "Crawlee", "image": "https://crawlee.dev/img/crawlee-og.png" },
    "twitter": { "card": "summary_large_image" },
    "headings": { "h1": ["Build reliable crawlers. Fast."], "h2": ["…"], "h3": [] },
    "schemaTypes": [],
    "images": { "total": 8, "withAlt": 6, "emptyAlt": 2, "missingAlt": 0, "altCoverage": 100 },
    "wordCount": 612,
    "links": { "internal": 42, "external": 11 },
    "issues": [
        { "severity": "info", "code": "SCHEMA_MISSING", "message": "No JSON-LD or microdata structured data found." }
    ],
    "seoScore": 98,
    "error": null,
    "scrapedAt": "2026-10-11T09:30:12.481Z"
}

How the score works

Every page starts at 100. Each issue subtracts points: error −15, warning −6, info −2 (minimum score 0). Typical errors: missing title, missing meta description, missing H1, noindex, very thin content, low alt-text coverage. Warnings: too long/short title or description, missing canonical, no viewport, multiple H1s. Info: missing Twitter card, no structured data, missing lang. Each issue has a stable code, so you can group and filter reliably.

A run summary (average score, issue counts by severity, top issue codes) is saved under STATS. Pages that return 404/410 or fail to load produce a record with an error and no score.

Pricing

Pay-per-event pricing:

Event	Price	When it is charged
page-audited	$0.005 (about $5 per 1,000 pages)	A page was fetched successfully and audited

Failed, blocked and 404 pages are not charged. Platform compute is included. Use the Console's maximum cost per run to cap spending.

Use cases

• Agency client reports: audit every URL in a sitemap export and ship a scored report.
• Pre-launch QA: catch missing titles, canonicals and noindex tags before go-live.
• Migration monitoring: re-run after a redesign to catch regressions.
• Competitor benchmarking: compare on-page SEO basics across rivals.
• Content audits: find thin pages and image alt gaps at scale.
• AI workflows: let an agent audit pages and suggest fixes.

FAQ

Does it crawl my whole site? No. It audits exactly the URLs you provide. Export your sitemap URLs and paste them in.

Does it render JavaScript? No. It audits the HTML the server returns, which is what most crawlers see first. Client-rendered tags added later will not appear.

Is the score the same as Google's ranking? No. It's an on-page hygiene indicator, not a ranking prediction. Use it to find and prioritise fixable basics.

Can I change the weights? Not currently. Every issue has a severity and code in the output, so you can re-score with your own weights downstream.

What about Core Web Vitals, backlinks or keyword rankings? Out of scope. This Actor focuses on on-page signals.

Why is my word count lower than expected? Navigation, header, footer and sidebars are excluded when the page has a clear main content area. Non-spaced languages (Chinese, Japanese) are counted less accurately.

MCP — use it from AI agents

Works with the Model Context Protocol, so Claude, Cursor and other MCP clients can audit pages on demand:

https://mcp.apify.com/?tools=fetch-actor-details,invincible_brookbed/seo-page-auditor

Try: "Audit these 10 landing pages and list the three most common issues."

Limitations

• HTML only: no JavaScript rendering, performance metrics or accessibility audits beyond alt text.
• No link-status checking (broken link detection) or duplicate-content comparison.
• Sites that block bots may fail; use proxies where needed.
• Scoring is heuristic and deliberately simple.