Static Cache Wrangler - Headless Assistant

No, this is a companion plugin that requires Static Cache Wrangler to be installed and active. It converts the HTML files that Static Cache Wrangler generates.

Commands:

• wp scw-headless info - Show system status and statistics
• wp scw-headless scan - List all cached HTML files ready for conversion
• wp scw-headless analyze <file> - Detect patterns in specific file
• wp scw-headless patterns - List all registered patterns with confidence scores
• wp scw-headless detectors - Show registered detector modules
• wp scw-headless convert --cms=sanity - Convert all files to Sanity format
• wp scw-headless targets - List available CMS targets All commands support --format=json for automation.

Current (Sanity conversion): The plugin creates a complete export package containing:

• data.ndjson - Sanity import data in newline-delimited JSON format
• asset-manifest.json - Asset references with URLs and metadata
• schemas/ - Sanity Studio schema definitions
• README.md - Import instructions for Sanity
• .zip archive - Complete package for download

The plugin uses a sophisticated multi-phase pipeline:

1. HTML Normalization - Strips WordPress-specific classes, IDs, and attributes while preserving semantic HTML
2. Pattern Detection - Uses XPath queries to find registered patterns (CSS selectors converted to XPath)
3. Priority Sorting - Processes patterns by priority (10=highest) to handle nested blocks correctly
4. Confidence Scoring - Each match includes confidence score (0.0-1.0) based on selector specificity
5. Content Extraction - Registered extractor functions parse matched DOM nodes into structured data
6. Conversion - Transform to target CMS format

Production Results (cachewrangler.com):

• Overall semantic conversion: 74%
• Core Gutenberg blocks: 100% accuracy (confidence: 1.00)
• Kadence Blocks: 90-95% accuracy (confidence: 0.90-0.98)
• 40 unique patterns registered
• 763 blocks processed across 15 pages
• 174 blocks fell back to raw HTML (navigation menus, advanced widgets) Per-Page Results:
• Simple pages: 82-86% conversion
• Mixed content pages: 74-82% conversion
• Complex pages (23 block types): 52% conversion

Falls back to rawHtml block type with pattern metadata. You can:

• Add custom pattern definitions via filters
• Report missing patterns on GitHub
• Wire up existing extractors (many already exist)

Yes! Three ways: 1. Register patterns via filter: php add_action('stcw_headless_patterns_loaded', function() { \STCW\Headless\Engine\Detector\PatternRegistry::register('my_block', [ 'selectors' => ['.my-block-class'], 'extractor' => [MyExtractor::class, 'extract_my_block'], 'priority' => 8, 'confidence' => 0.95, ]); }); 2. Create detector modules (for larger block libraries) 3. Use pattern inheritance: php // Extend existing patterns PatternRegistry::register('custom_button', [ 'extends' => 'button', // Inherits base button selectors 'selectors' => ['.my-custom-button'], // Adds custom selectors ]);

Yes! The plugin is designed with a pluggable architecture: php // Register custom CMS target add_action('stcw_headless_register_targets', function() { $my_target = new My_CMS_Target(); \STCW\Headless\Engine\Target\TargetRegistry::register($my_target); }); Your target class must implement TargetInterface with methods for convert(), generate_schemas(), and export().

By default, these paths are excluded:

• assets/ - Static assets (CSS, JS, images)
• author/ - Author archives
• category/, tag/ - Taxonomy archives
• index.php/ - WordPress quirks
• feed/, wp-json/ - API endpoints
• sitemap/, 404/ - Utility pages
• Blog index pages (Posts Page in Settings → Reading) Filter via stcw_headless_excluded_paths to customize.

WordPress "Posts Page" archives (the page set as your blog index in Settings → Reading) are intentionally skipped because they contain dynamic post loops, not static content. Individual blog posts are converted successfully. To recreate your blog index in Sanity:

1. Import individual posts (automatically converted)
2. Use this GROQ query to fetch posts: groq *[_type == "post"] | order(publishedAt desc) { title, slug, excerpt, publishedAt }
3. Build your blog index view in your frontend

Kadence Blocks is currently supported. Support is being considered for Elementor, Otter Blocks, Divi, and more.

v2.1.0 Benchmarks (cachewrangler.com test site):

• 15 pages converted in ~6 seconds
• Small page (10 KB): ~0.2 seconds
• Medium page (50 KB): ~0.5 seconds
• Large page (100 KB): ~1.0 seconds
• Batch (100 pages): ~45 seconds
• Pattern detection: XPath-based (efficient)
• Memory: ~20 MB per page

Yes! Use wp scw-headless analyze <file> to see:

• Patterns detected
• Confidence scores
• Asset references
• Potential issues
• Extraction preview Example:

File: index.html (104 KB) Patterns Found: 71 paragraph 20 Confidence: 1.00 heading 15 Confidence: 1.00 separator 14 Confidence: 1.00 kadence_button 6 Confidence: 0.90 kadence_accordion 2 Confidence: 0.95 ... Confidence Distribution: High (≥0.95): 60 Medium (0.85+): 11 Low (<0.85): 0 ```

• Free users: GitHub Issues
• Documentation: wp2headless.com/docs

Static Cache Wrangler may create gzipped files or use compression. The plugin handles this automatically by reading the actual index.html files within cached directories.