Free SEO audit (seo:audit)

php artisan seo:audit answers one question in one command, for free: what's wrong with my SEO right now? It iterates your HasSEO models in-process — no queue, no license, no network — and prints a per-page pass / warn / fail table with a summary.

php artisan seo:audit

+--------------+--------+----------------------------+
| Page         | Status | Findings                   |
+--------------+--------+----------------------------+
| Post #1      | PASS   | —                          |
| Post #2      | WARN   | notice title_too_short     |
| Post #3      | FAIL   | critical missing_title     |
|              |        | warning missing_description|
+--------------+--------+----------------------------+

3 page(s) — 1 passed, 1 warned, 1 failed
4 issue(s) — 1 critical, 1 warning, 2 notice

What it checks

The audit runs only the metadata execution class — the checks that are resolvable from the model and the resolver alone, with no page fetch:

Check	Codes
Title / description present (fallback-aware)	missing_title, missing_description
OG image present (fallback-aware)	missing_og_image
Title / description length	title_too_long, title_too_short, description_too_long, description_too_short
Duplicate title / description across the site	duplicate_title, duplicate_description
Robots conflicts & suspicious noindex	robots_conflict_indexing, robots_conflict_following, noindex_warning
Canonical format / cross-domain / shared / insecure	invalid_canonical, cross_domain_canonical, shared_canonical, insecure_canonical
Answer-readiness (AEO) — article structured data	aeo_missing_author, aeo_article_missing_date
Focus keyword set (opt-in)	missing_focus_keyword

These are the same issue codes the Pro scan emits, so a finding means the same thing here as it does there. Length reuses the editor's 60/160 thresholds (measured against the resolved value, suffix included), so the audit never contradicts the character counters in the Filament editor.

The answer-readiness (AEO) checks fire only when a page declares article-type JSON-LD (Article, BlogPosting, NewsArticle, …) that is missing a signal AI answer engines rely on — an author entity for attribution and E-E-A-T, or a datePublished / dateModified for recency. A page without an article is never flagged, so the audit stays quiet where AEO doesn't apply. They are advisory (notice-level) and held out of the Pro 0–100 score.

What it does not check — the capability boundary

A free in-process audit can never equal the full Pro scan, and the command says so on every run. It does not run:

• Rendered-HTML checks — missing_h1, multiple_h1, missing_image_alt, thin_content, mixed_content. These need the page's served HTML.
• Live-canonical network checks — canonical_target_broken / _redirect / _noindex. These need an outbound (guarded) fetch.
• The numerical 0–100 score. The score is a Pro feature, persisted on the scan-result record with a versioned rubric — see SEO score.

Those ship in the Pro scan — see the full issue registry.

Choosing what to audit

By default the command audits the models listed under seo.audit.models, falling back to seo.sitemap.models:

// config/seo.php
'audit' => [
    'models' => [
        \App\Models\Post::class,
        \App\Models\Page::class,
    ],
],

Or pass models explicitly:

php artisan seo:audit --model="App\Models\Post" --model="App\Models\Page"

Options

Option	Effect
--model=	A HasSEO model class to audit (repeatable). Overrides config.
--locale=	Resolve SEO data in this locale (defaults to the app locale).
--limit=	Max records to audit per model (0 = all).
--issues-only	Only list pages that have at least one issue.
--strict	Exit with a non-zero status when any issue is found — for CI.
--json	Emit machine-readable JSON (pages, summary, coverage) instead of the table.

CI gate

--strict turns the audit into a build check:

php artisan seo:audit --strict

It exits 1 if any page warns or fails, 0 when every audited page passes.

JSON

php artisan seo:audit --json

{
  "pages": [
    {
      "model": "App\\Models\\Post",
      "key": 3,
      "label": "Post #3",
      "url": "https://example.com/blog/...",
      "status": "fail",
      "issues": [
        { "code": "missing_title", "severity": "critical", "field": "title", "message": "Page is missing a title tag." }
      ]
    }
  ],
  "summary": { "pages": 3, "passed": 1, "warned": 1, "failed": 1, "issues": 4, "by_severity": { "critical": 1, "warning": 1, "notice": 2 } },
  "skipped": [],
  "coverage": { "executes": "metadata", "note": "...", "reference": "https://rankbeam.dev/pro/scan-issues" }
}

Focus keywords

The missing_focus_keyword notice is off by default. It only fires once you opt into the focus-keyword workflow:

// config/seo.php
'keywords' => [
    'enabled' => true,
],

The Pro scan reads the same flag, so the audit, the scan, and the Pro editor nag always agree. Set a page's keywords with the Filament focus-keyword field or $model->saveSEO(['focus_keywords' => [['keyword' => 'laravel seo', 'is_primary' => true]]]).

When a value isn't what you expect: seo:explain

seo:audit tells you what's wrong; seo:explain tells you why a field resolved the way it did — which layer (config / default / computed / explicit) set each value, what it overrode, and what post-processing (title suffix, canonical strip, indexing guard) changed afterwards. Reach for it when an audit finding or a rendered tag is surprising:

php artisan seo:explain "App\Models\Post" 42