Arabic & Multilingual Website Auditor

Audit public HTML pages for Arabic language declarations, document direction, metadata, canonical and hreflang signals. Designed for agencies maintaining Arabic and translated websites, with Arabic explanations and machine-readable exports.

What you receive

One dataset record for each attempted page: URL, HTTP status, title, description, language, direction, canonical, hreflang declarations, robots signals, H1 headings, image-alt counts and actionable findings. The key-value store contains REPORT (JSON), REPORT-CSV and REPORT-HTML (Arabic report).

Use the input form or pass:

{"startUrls":[{"url":"https://example.com/"}],"maxPages":10,"crawl":true,"delayMs":500,"timeoutSeconds":15}

Maximum 20 starting URLs and 200 page attempts per run. Requests stay on each starting URL's origin. Use the final canonical origin as input, for example https://www.example.com/ rather than an origin that redirects there. Discovered query-string URLs and common downloads are skipped. Submitted query-string URLs are supported.

Checks

• HTTP errors and redirect chains.
• Missing or multiple titles, missing description, H1 structure review.
• Missing/invalid document language and predominantly Arabic content with a conflicting language declaration.
• Arabic document direction review (HTML dir only).
• Robots noindex signals (including Googlebot-specific declarations, exposed as signals to review).
• Canonical presence, validity and differences from the fetched URL.
• Hreflang syntax, absolute URLs, duplicate language conflicts and missing self-reference.
• Images missing the alt attribute; empty alt attributes are allowed for decoration.

Limits and safe operation

This is an HTTP HTML audit, not a rendered-browser audit. Only UTF-8 HTML is supported. JavaScript content, CSS direction, hreflang reciprocity, alternate-target health, HTTP Link canonical and actual search-engine index status are not checked. No SEO ranking score is invented. An intentional noindex or alternate canonical is not automatically an error.

The crawler respects robots.txt. A robots outage or rate limit fails closed; 401/403 disallows crawling. Nonstandard ports, credentials in URLs and private/reserved network addresses are blocked. DNS answers are checked and the socket is pinned to a checked address. Each response has a size limit, a request timeout and a redirect limit. No login, proxy, CAPTCHA bypass or LLM API is used. Crawl delays over 10 seconds cause the origin to be skipped.

Pricing

Use pay-per-event with one custom event page-audited, a charge for one successfully fetched HTML page (HTTP 2xx/3xx). Remove or zero the synthetic apify-actor-start and apify-default-dataset-item events. Failed requests, robots-blocked URLs and non-HTML responses are diagnostics without custom event charges. The runtime rejects conflicting event configuration and stops before a new page when the remaining event budget is insufficient.

$0.01 per successfully audited HTML page ($10 per 1,000), including platform usage. HTTP errors, robots-blocked URLs, duplicate final URLs and non-HTML responses have no custom event charge. Set a maximum run budget to control spending; the minimum supported budget is $0.01. The Actor respects the remaining page-event budget before requesting another page.

Source: https://docs.apify.com/actors/publishing/monetize and https://docs.apify.com/sdk/js/docs/concepts/pay-per-event

Local development

Requires Node.js 22+.

npm ci
npm test
npm run audit -- example-input.json results
npm start

audit runs without an Apify account and writes local JSON, CSV and HTML. start uses the Apify SDK storage lifecycle; provide input through storage/key_value_stores/default/INPUT.json locally or the input form on Apify. Local tests do not charge a customer. Production event charging must be tested after the event is configured in Console; it is not activated by these files alone.

Deployment

Deploy this folder as an Actor using the official Apify CLI or create an Actor with these source files in the Console. The Docker image builds with npm ci --omit=dev. Build and run privately, inspect the dataset and reports, then measure resource costs. Fill developer/payout requirements personally in Console. Publish only after a successful cloud run and an approved price. Routine maintenance includes dependency updates, checking failures and customer support.