Website SEO & Metadata Extractor

What does Website SEO & Metadata Extractor do?

Run batch on-page SEO checks for up to 50 public HTML pages and get structured metadata plus prioritized review suggestions. Inspect titles, descriptions, headings, canonical links, images and internal/external links. Try https://apify.com/ as a small first run. Apify provides API access, saved tasks, schedules, integrations and run monitoring.

Why use this SEO audit?

Agencies can review client landing pages, developers can check metadata after a deployment, and content teams can compare titles and descriptions. Each issue includes evidence and a recommendation. The summary identifies exact duplicate titles and descriptions among processed pages. These checks support human review; they are not a ranking score or a guarantee of search performance.

How to use it

1. Enter a single URL or add explicit Batch page URLs.
2. Set Maximum pages. Start with two pages.
3. Start the Actor and review Page audits.
4. Open Batch summary and errors for skipped URLs, duplicates and stopping reason.
5. Download results or connect your workflow using the Apify API.

Input

Batch URLs take precedence over the single URL. Duplicate URLs and fragments are removed. Links are not automatically crawled.

{"startUrls":[{"url":"https://apify.com/"},{"url":"https://apify.com/store"}],"maxPages":2,"maxLinks":25,"maxImages":20,"timeoutSeconds":20}

Existing integrations using url remain supported. maxPages defaults to 20 (1–50), maxLinks to 100 (0–1000), maxImages to 50 (0–500), and timeoutSeconds to 20 (5–60). Empty input audits the Apify homepage.

Output

One dataset record is written per successfully analyzed HTML page. Failed, blocked and non-HTML pages appear in SUMMARY instead. Download the dataset in JSON, CSV or Excel.

Illustrative fields:

{"requestedUrl":"https://example.com/","title":"Example Domain","success":true,"issues":[{"code":"missing_description","severity":"medium","evidence":"No meta description","recommendation":"Write a page-specific summary for search snippets."}]}

Data table

Field	Meaning
requestedUrl / finalUrl	Input URL and final destination
statusCode / contentType	HTTP response information
title / metaDescription / canonicalUrl	Search metadata
headings / h1Count	H1–H6 structure
wordCount	Whitespace-separated body words, excluding scripts/styles/templates
links / linkSummary	Capped details and uncapped counts
images / imageCount	Capped details and total count
seoChecks / imagesEmptyAlt	Missing alt attributes and separately counted empty alt text
issues / recommendations	Review priorities, evidence and fixes
linksTruncated / imagesTruncated	Whether lists were capped

SUMMARY contains attempted pages, saved results, errors, duplicate metadata, duration and stopReason. Counts may be partial when spending or time limits stop a run.

Pricing and cost estimation

See Pricing for current event prices. At the current base rate, 100 analyzed pages cost $0.40 in result events, plus Actor-start events. Failed pages do not create result events; the start event still applies. Batch pages to reduce repeated startup overhead. Set your run spending limit and maxPages. The Actor stops when the SDK reports the result-charge limit.

Tips and limitations

Server-rendered HTML only. JavaScript-only content, logged-in pages, PDFs, browser performance metrics and backlink analysis are outside scope. Links are extracted, not fetched for broken-link testing. Whitespace-based word counts are not linguistic segmentation for every language. Empty alt may be intentional for decorative images.

Requests respect robots.txt, with a minimum one-second page delay and bounded retries for selected transient HTTP errors. Unverifiable robots responses cause skips. Limits include 2 MB per response, five redirects and a roughly four-minute work budget checked between pages. Only public HTTP(S) URLs on standard ports are supported.

FAQ and support

Use URLs you are permitted to access and respect site terms. Review noindex flags and multiple H1s in context. Missing data can indicate client-side rendering. For issues or custom workflow requests, use the Issues tab with a public URL and run ID, without passwords or private data.