Scan Website

The scan command fetches a public URL, measures the total page size, and audits the page's i18n and SEO health. It produces a scored report (0–100) covering HTML attributes, canonical links, hreflang tags and their return links, robots.txt, sitemaps, localised internal links, and JavaScript bundle locale weight.

It also reports how the site encodes the locale in its URLs (routing strategy) and which framework, i18n library, translation management system (TMS) or translation proxy it uses. The same checks power the online i18n SEO scanner and the Intlayer Chrome extension.

No extra dependencies are required. When puppeteer is installed the scan can capture lazy-loaded JavaScript chunks for a more precise bundle analysis; otherwise it falls back to inspecting eagerly-loaded scripts declared in the HTML.

Usage

npx intlayer scan <url>

Example

npx intlayer scan https://example.com

Sample output:

🔍 Scanned https://example.com/fr (basic mode)

Score: 91/100
Page size: 10.60 MB (HTML 42.31 KB)
Locales: fr, en, es, de
Routing: locale prefix except for the default locale (every hreflang alternate but "en" starts with a locale segment, default locale: en)

Stack:
  Framework Next.js 15.1.0
  i18n library next-intl
  TMS Crowdin
Checks:
  ✓ html lang attribute
  ✓ html dir attribute
  ✓ locale signals consistent (lang, URL, hreflang)
  ⚠ og:locale meta tag
      Missing <meta property="og:locale">: social previews default to en_US
  ✓ canonical link
  ✓ hreflang tags
  ✓ x-default hreflang
  ✓ hreflang alternates link back
  ✓ unused bundle locale content
  ✓ localized internal links
  ⚠ all internal links keep the locale
      2 internal links leave the "fr" locale (0 to another locale, 2 without locale)
        <a href="/pricing">Tarifs</a>
  ✓ crawlable language switcher
  ✓ robots.txt present
  ✓ robots.txt keeps localized URLs crawlable
  ✓ sitemap present
  ✓ sitemap lists every locale
  ✓ sitemap has alternate links
  ✓ sitemap has x-default

Bundle locale weight:
  Translations shipped: 120.50 KB
  Unused (other locales): 45.20 KB (37%)

Options

<url> (required)

The fully-qualified URL to scan (e.g. https://example.com).

--no-deep

Disable the deeper render-based scan.

By default the command attempts to use puppeteer to render the page in a headless browser, capture lazy-loaded JavaScript chunks, and measure the true wire-transfer size. If puppeteer is not installed, the command automatically falls back to basic mode.

Pass --no-deep to force basic mode even when puppeteer is available.

Example: npx intlayer scan https://example.com --no-deep

--json

Output the full scan result as a JSON object instead of a formatted report. Useful for programmatic consumption or CI pipelines.

Example: npx intlayer scan https://example.com --json

Standard configuration options

• --base-dir: Base directory used to locate the intlayer.config.* file.
• -e, --env: Target environment (e.g. development, production).
• --env-file: Path to a custom .env file.
• --no-cache: Disable configuration cache.
• --ci: Run the command in every Intlayer project of the monorepo (or only the current one when run from a project directory). Per-project credentials can be injected via INTLAYER_PROJECT_CREDENTIALS, a JSON map of project path to { "clientId", "clientSecret" }.
• --verbose: Enable verbose logging (default in CLI mode).
• --prefix: Custom log prefix.

Routing strategy

The locale pattern shared by the page's hreflang alternates reveals how the site routes its locales. Without alternates, the scanned URL alone is used (low confidence).

Link, canonical, robots.txt and sitemap checks read every URL through this strategy. For example, an unprefixed link is correct on the default locale of a prefix-no-default site, and a link without ?lang= leaves the locale on a search-params site.

Detected stack

Frameworks, i18n libraries (Intlayer, i18next, react-i18next, next-i18next, next-intl, use-intl, react-intl, vue-i18n, @nuxtjs/i18n, Lingui, svelte-i18n, Paraglide, ngx-translate, Transloco, Polylang, WPML…), translation management systems (Crowdin, Phrase, Lokalise, locize, Transifex, Tolgee, Localazy, SimpleLocalize, Localizely, Smartling, Intlayer CMS) and translation proxies (Weglot, Localize, GTranslate…) are identified from the HTML, the loaded resources and the JavaScript bundles. Deep mode also reads window globals and cookies.

What is checked

A warning earns half of the weight. The final score is the weighted sum of the checks that ran, expressed as a percentage (0–100). Failing checks print the first issues found; use --json for the full details.

Using the scan function programmatically

The scan function is also exported from @intlayer/cli so it can be called from your own scripts:

import { scan } from "@intlayer/cli";

await scan("https://example.com", {
  deep: false,
  json: false,
});

For lower-level access, scanWebsite from @intlayer/engine/scan returns a structured ScanResult object:

import { scanWebsite } from "@intlayer/engine/scan";

const result = await scanWebsite("https://example.com", { deep: false });
console.log(result.score, result.totalPageSize, result.events);