Real data or a clear unavailable state
Search Console, GA4, PageSpeed and provider-backed results are only displayed when fetched. Missing data stays unavailable—not replaced with plausible defaults.
Portfolio/Case study
WordPress SEO platform · AI-nativeMagePeople
SEO Suite
Every search signal, one source of truth.
A full WordPress SEO system for metadata, schema, crawling, content health, analytics, WooCommerce, and assistant-driven workflows—designed to use real site and provider data, never invented numbers.
Commercial project · source repository is private35+Assistant SEO tools
7Content-health checks
4SEO plugin importers
1Authoritative schema graph
The problem
SEO plugins often disagree with themselves
A site can emit one canonical URL in the head, submit another in its sitemap, generate a third in schema, and still show analytics-looking numbers that never came from an analytics provider.
One authoritative output
Metadata, canonical resolution, indexability, Open Graph, Twitter cards, JSON-LD and sitemaps share the same resolved content state.
Safe automation by default
Every assistant action passes WordPress capabilities. External interfaces start off, imports support dry runs, and destructive changes require explicit intent.
Platform surface
From a single page title to the whole search graph
The modules share resolvers and indexes rather than behaving like unrelated mini-plugins.
On-page SEO
Titles, descriptions, focus keywords, social previews, robots directives and canonical controls for posts, terms and archives.
Structured-data graph
A deduplicated JSON-LD graph for WebSite, Organization, Person, Article, Breadcrumbs, products and custom schema.
Search intelligence
Google Search Console, GA4, PageSpeed Insights, URL inspection, rank tracking and traffic-recovery workflows.
Crawling & indexing
XML, news, video, RSS and HTML sitemaps; robots and file tools; redirect monitoring; broken-link and orphan-content discovery.
Content health
Duplicate metadata, cannibalization, thin or stale pages, invalid canonicals and incomplete author identity in one bounded audit.
WooCommerce SEO
Variable-product ProductGroup markup, identifiers, offer availability, merchant return policies and store-aware product data.
AI workflows
35+ focused tools for auditing, drafting, updating, inspecting, comparing and reporting through secure WordPress interfaces.
Builder compatibility
Stored-content resolution for Gutenberg, Classic Editor, Elementor, Divi, Oxygen and Breakdance without rendering unsafe shortcodes.
Architecture
Resolve once, publish everywhere
The same indexability and content resolution pipeline feeds every public and administrative surface.
WordPress sources
- Posts, terms & users
- SEO metadata
- Builder documents
- WooCommerce products
- Imported plugin data
Shared resolution layer
- Effective canonical URL
- Indexability decision
- Resolved visible content
- Entity and identity graph
- Cached audit indexes
Consistent outputs
- Head metadata & JSON-LD
- XML and HTML sitemaps
- Content-health reports
- REST and assistant tools
- Admin recommendations
Content health
Seven checks that explain the problem
Each finding links back to a real WordPress object and states why it was flagged.
- Duplicate SEO titles
- Duplicate meta descriptions
- Focus-keyword cannibalization
- Thin resolved content
- Stale published content
- Invalid or chained canonicals
- Missing author biographies
Real-data contract
Nothing is silently fabricated
The interface distinguishes site-derived data, provider data, cached data and unavailable data.
WordPress contentLive site query
Search ConsoleGoogle API or unavailable
Google Analytics 4Google API or unavailable
PageSpeedProvider response or error state
Rank observationsRecorded checks, never estimates
Canonical validityResolved against real objects
WooCommerce
Product schema follows the actual catalog
Variable products become a connected ProductGroup with real variants, offers and identifiers—not a flattened approximation.
ProductGroup · parent product
Product · Blue / Medium · offer
Product · Blue / Large · offer
Product · Green / Medium · offer
Rich data, guarded output
Prices, currency, availability, images, attributes, SKU, GTIN, MPN and ISBN are read from WooCommerce objects. Merchant return policy markup remains disabled until the store provides the required policy configuration.
- Variants link back to their group
- Offers use live stock and pricing
- Optional policy data fails closed
Assistant layer
35+ tools, with WordPress still in control
AI clients can inspect and improve SEO without bypassing the site's permission model.
Tool access uses WordPress capabilities and authenticated interfaces. OAuth 2.1 with PKCE is supported, with Application Passwords as a WordPress-native fallback. REST and assistant access are disabled until an administrator enables them.
- Audit posts, terms, products and archives
- Read provider-backed analytics and search data
- Draft metadata and schema changes
- Run imports in dry-run mode first
- Compare current and proposed values
- Reject anonymous and under-privileged requests
Verification
Built as a system, tested as a system
The final audit checked syntax, public output, permissions, schema integrity and representative live WordPress fixtures.
441PHP files passed syntax linting
28JavaScript files passed syntax checks
1 + 1Single canonical and single JSON-LD graph
401 / 200Anonymous denied, authorized admin accepted
2Real variants verified in ProductGroup fixture
7 / 7Content-health checks return live site findings
Canonical consistency is a shared service
Head output, schema and sitemaps ask one resolver instead of reproducing URL logic in three modules.
Stored builder data is inspected safely
Audits understand supported builders while avoiding front-end shortcode execution and side effects.
Expensive indexes are invalidated, not rebuilt blindly
Content changes mark cached results stale; scheduled or requested work refreshes them within bounded batches.
Compatibility targets real hosting
The codebase maintains a PHP 7.4 floor and WordPress 6.0 baseline rather than assuming a modern managed stack.
Under the hood
Technology & integrations
WordPress platform
PHP 7.4+WordPress 6.0+REST APIAbilities APIWP-CronWP-CLIWooCommerceMultisite
Search & analytics
Search ConsoleGA4PageSpeed InsightsURL InspectionSchema.orgOpen GraphXML Sitemaps
Editors & builders
GutenbergClassic EditorElementorDiviOxygenBreakdance
Migration & security
Rank MathYoast SEOAIOSEOSEOPressOAuth 2.1 + PKCEApplication PasswordsCapability gates
Want the architecture walkthrough?
The source repository is private. I can walk through the data model, indexing pipeline, schema graph, integrations, and safe assistant layer.