Explain the resolution (seo:explain)
Rankbeam resolves a page's SEO through a layered precedence chain — config, database defaults (global / model-type / route), computed model values, then explicit seo_meta — followed by post-processing (title suffix, canonical, image absolutization) and the indexing guard. When the rendered <title> or robots tag isn't what you expected, seo:explain shows you exactly which layer set each field, and what it overrode.
It's read-only, needs no network or license, and never re-implements the merge: the attribution comes from the resolver's own layer contributions and the final values from the real resolver — so the explanation can't drift from what actually renders.
Usage
# Explain a specific record
php artisan seo:explain "App\Models\Post" 42
# Explain the first record of a model
php artisan seo:explain "App\Models\Post"
# With a route-defaults layer and a locale
php artisan seo:explain "App\Models\Post" 42 --route=posts.show --locale=de
# Machine-readable
php artisan seo:explain "App\Models\Post" 42 --jsonThe model must use the HasSEO trait.
Reading the output
SEO resolution — Post #42 (locale: en, route: posts.show)
Layers, low → high: config · global · model-type · route · computed · explicit
Field Final value Set by Overrode
title My Post | Acme computed —
↳ title suffix ' | Acme' appended
description A hand-written summary… explicit computed: "An auto excerpt…"
canonical https://acme.com/blog/my-post post-processing —
↳ derived from model getUrlForSEO() (query string stripped)
robots noindex,nofollow explicit config: index,follow
↳ indexing guard forced 'noindex,nofollow' (environment 'staging' …)
og_image https://acme.com/share.jpg explicit config: /default-og.jpg
↳ absolutized from '/share.jpg'- Set by — the winning layer (the highest-precedence layer that set a non-null value), or
post-processingwhen no layer set the field but a value was derived (a canonical from the request/model URL, an og:url from the canonical, an absolutized image). - Overrode — every lower-precedence layer that offered a value and lost, in order, so you can see what was shadowed.
- ↳ notes — post-processing that changed the value after the layer merge: the title suffix, the canonical query-string strip, og:url derivation, image absolutization, and the indexing guard forcing
noindexabove every layer.
og:type and twitter:card
These two carry non-null framework defaults (website / summary_large_image), so the highest layer that sets them — normally computed — wins over config. A page with no stored seo_meta row contributes nothing for them, so a computed og:type like article is never shadowed by a bare website. That's faithful to how the merge actually resolves them.
Site-level resolution
Per the amendment for the site-config ledger, seo:explain also reports the site-wide values whose source is a frequent source of confusion — which source set the canonical host, the site name, and the default locale:
Site-level resolution
Value Resolved Source
Site name Acme env (APP_NAME)
Default locale en config (app.locale)
Canonical host acme.com programmatic (model getUrlForSEO())The canonical host is the value most worth checking — a wrong host (a leaked localhost, an http:// on an https site, the app URL not matching the model URL) is a classic cause of self-canonical bugs.
JSON output
--json emits the full trace — target, per-field winner / losers / final / notes, and the site_level ledger — for tooling or CI:
{
"target": { "model": "App\\Models\\Post", "id": 42, "route": "posts.show", "locale": "en" },
"fields": {
"title": {
"final": "My Post | Acme",
"winner": { "layer": "computed", "value": "My Post" },
"losers": [],
"notes": ["title suffix ' | Acme' appended"]
}
},
"site_level": {
"canonical_host": { "value": "acme.com", "source": "programmatic (model getUrlForSEO())" }
}
}See also
- Resolver precedence — the full chain
seo:explaintraces. - Free SEO audit —
seo:auditfinds what's wrong;seo:explainshows why a value is what it is.