Ranklint

Lint your SEO before Google does.

SEO toolkit for Nuxt 4: sitemap/robots/JSON-LD generation out of the box, an SEO linter in dev via DevTools, and regression control in CI.

• Runtime — the module serves sitemap.xml, robots.txt, and typed JSON-LD by itself
• Dev — an "SEO" tab in Nuxt DevTools: heading outline, meta, JSON-LD with validation, page audit
• CI — the CLI crawls a deployed site and fails on thresholds; junit/markdown/json reports

Quick start

Module

npx nuxi module add @ranklint/nuxt

Works with zero configuration: /sitemap.xml (static routes from app/pages/) and /robots.txt (non-prod environments are closed to indexing) are available right away.

// nuxt.config.ts
export default defineNuxtConfig({
  modules: ['@ranklint/nuxt'],
  ranklint: {
    site: { url: 'https://example.com', name: 'Example' },
    sitemap: {
      sources: [
        '/api/seo/urls',
        async () => (await $fetch<string[]>('/api/products')).map(slug => ({ loc: `/p/${slug}` })),
      ],
      cacheTtl: 3600,
    },
    robots: { mode: 'owner' },
    jsonLd: true,
  },
})

Every block can be disabled with false — disabled code is not registered at all.

A function source is serialized into the server bundle via toString(), so it must be self-contained: Nitro globals ($fetch) are available, but closures over variables and imports from nuxt.config are not.

<script setup>
useJsonLd('Product', {
  name: 'Widget',
  offers: { price: 9.99, priceCurrency: 'USD' },
})

useRanklintIgnore(['headings:single-h1'])
</script>

In dev, JSON-LD is validated against Schema.org schemas (Product, Article, BreadcrumbList, Organization, WebSite) with console warnings; in prod the validator is completely tree-shaken out of the bundle. The module's client runtime is under 1 KB gzip.

CLI

npm i -D @ranklint/cli
npx playwright install chromium

ranklint audit --url https://uat.example.com          # audit a live site
ranklint audit --start .output/server/index.mjs        # or self-contained from a build

Exit code 1 when errors are found. Reporters: markdown (default), json, junit (--reporter, --output).

Rules are configured in ranklint.config.ts with ESLint semantics:

import { defineRanklintConfig } from '@ranklint/core'

export default defineRanklintConfig({
  site: { url: 'https://example.com' },
  rules: {
    'meta:title-length': ['error', { min: 30, max: 60 }],
    'meta:description-length': 'off',
  },
  crawl: { concurrency: 5, maxPages: 2000, ignore: ['/admin/**'] },
})

GitLab CI

include:
  - remote: 'https://raw.githubusercontent.com/ranklint/ranklint/main/presets/gitlab-ci/seo.yml'

seo:audit:
  extends: .ranklint-audit
  needs: [deploy:uat]
  variables:
    RANKLINT_URL: $CI_ENVIRONMENT_URL

ranklint diff --base main resolves the base report from CI artifacts: in GitLab — from the branch's job artifact, in GitHub Actions — from the ranklint-report artifact (zip extraction is built in). A missing base is not an error: diff degrades to a full report.

Rules

42 rules across the meta, headings, canonical, links, i18n, structured-data, images, robots, indexability, and http categories — see the full reference in docs/rules.md.

Every rule accepts 'error' | 'warn' | 'off' or [severity, options]. Inline suppression on a page — useRanklintIgnore([...]).

For ranklint.config.json/.jsonc there is a JSON schema with autocompletion — schemas/ranklint-config.schema.json. In a TS config, defineRanklintConfig provides the same.

Also included: SEO diff between branches (ranklint diff), multi-app zones, per-route Lighthouse thresholds, watch mode, production monitoring with alerts (slack/telegram), CrUX and Search Console data, crawl-budget analysis, and custom rules.

Custom rules

// ranklint.config.ts
import { defineCheck } from '@ranklint/checks'
import { defineRanklintConfig } from '@ranklint/core'

export default defineRanklintConfig({
  site: { url: 'https://example.com' },
  customChecks: [
    defineCheck({
      id: 'myteam:no-lorem',
      category: 'meta',
      severity: 'warn',
      scope: 'page',
      docs: 'https://wiki.myteam.dev/seo/no-lorem',
      async run(ctx) {
        const text = ctx.document?.body?.textContent ?? ''
        if (!/lorem ipsum/i.test(text)) return []
        return [{
          checkId: 'myteam:no-lorem',
          severity: 'warn',
          message: 'Placeholder text found',
          url: ctx.page!.url,
          suggestion: 'Replace lorem ipsum with real content',
        }]
      },
    }),
  ],
  rules: {
    'myteam:no-lorem': 'error',
  },
})

Custom rules are first-class alongside built-ins: configured via rules, disabled with 'off' and useRanklintIgnore. The CheckContext contract is stable within a major version.

Presets are plugged in via extends (native c12): @ranklint/preset-default pins all built-in rules at their default severities. Earlier layers take precedence over later ones, and ranklint.config itself overrides all layers:

export default defineRanklintConfig({
  extends: ['./agency-preset.ts', '@ranklint/preset-default'],
  site: { url: 'https://example.com' },
  rules: { 'images:alt-required': 'error' },
})

Packages

| Package | Purpose | | --- | --- | | @ranklint/nuxt | Nuxt 4 module: sitemap, robots, useJsonLd, DevTools tab | | @ranklint/cli | ranklint audit — crawler (Playwright) + checks + reports | | @ranklint/core | Engine: crawler, runner, config (no Nuxt dependencies) | | @ranklint/checks | Rules + Schema.org schemas | | @ranklint/reporters | markdown / junit / json | | @ranklint/devtools | Vue panel for Nuxt DevTools: live checks of the current page | | @ranklint/preset-default | Preset with the built-in rules for extends |

Requirements: Nuxt ^4.0.0, Node.js >= 20. MIT license.