Content type selection for RelevantFlow \u2014 index Posts, Pages, Products, or any custom post type. New guided setup wizard after activation. Fixes AltGenius Pro link.<\/p>","1.0.4":"
Full SQLite compatibility \u2014 similar posts table now works on WordPress Playground and SQLite-based environments. Existing MySQL installs auto-migrated.<\/p>","1.0.3":"
Fixes JavaScript errors on WordPress Playground and missing script dependencies for AltGenius editor\/media upload.<\/p>","1.0.2":"
Fixes admin UI loading issue on WordPress Playground and older WordPress versions.<\/p>"},"ratings":[],"assets_icons":{"icon-128x128.png":{"filename":"icon-128x128.png","revision":3483604,"resolution":"128x128","location":"assets","locale":""},"icon-256x256.png":{"filename":"icon-256x256.png","revision":3483604,"resolution":"256x256","location":"assets","locale":""}},"assets_banners":{"banner-1544x500.jpg":{"filename":"banner-1544x500.jpg","revision":3483604,"resolution":"1544x500","location":"assets","locale":""},"banner-772x250.jpg":{"filename":"banner-772x250.jpg","revision":3483604,"resolution":"772x250","location":"assets","locale":""}},"assets_blueprints":{"blueprint.json":{"filename":"blueprint.json","revision":3484521,"resolution":false,"location":"assets","locale":"","contents":"{\"$schema\":\"https:\\\/\\\/playground.wordpress.net\\\/blueprint-schema.json\",\"description\":\"Preview the Smart Local AI plugin \\u2014 browser-based private AI tools for WordPress.\",\"landingPage\":\"\\\/wp-admin\\\/admin.php?page=smart-local-ai\",\"preferredVersions\":{\"php\":\"8.2\",\"wp\":\"6.9\"},\"steps\":[{\"step\":\"login\",\"username\":\"admin\",\"password\":\"password\"},{\"step\":\"installPlugin\",\"options\":{\"activate\":true},\"pluginData\":{\"resource\":\"wordpress.org\\\/plugins\",\"slug\":\"smart-local-ai\"}},{\"step\":\"writeFile\",\"path\":\"\\\/wordpress\\\/wp-content\\\/mu-plugins\\\/smart-local-ai-demo.php\",\"data\":\" 'Getting Started with Machine Learning in the Browser', 'content' => 'Machine learning is transforming how we build software. From recommendation engines to image recognition, ML models are becoming essential tools for modern developers. Browser-based inference using WebAssembly and WebGPU is making AI accessible to everyone without cloud dependencies or API costs. Learn how client-side ML works and why it matters for privacy.', 'excerpt' => 'An introduction to browser-based machine learning and why it matters.', 'cat' => $ai),\\n\\t\\tarray('title' => 'Privacy-First AI: Why Client-Side Processing Matters', 'content' => 'As data privacy regulations like GDPR and CCPA tighten worldwide, the demand for privacy-first AI solutions is growing rapidly. Client-side machine learning allows websites to offer intelligent features without sending user data to external servers. This approach ensures compliance while delivering powerful AI capabilities directly in the browser.', 'excerpt' => 'How client-side ML is reshaping privacy-compliant AI tools.', 'cat' => $ai),\\n\\t\\tarray('title' => 'Understanding Semantic Search and Vector Embeddings', 'content' => 'Semantic search goes beyond keyword matching to understand the meaning behind queries. Using vector embeddings and cosine similarity algorithms, semantic search delivers more relevant results by understanding context and intent. This technology powers modern recommendation engines and content discovery systems.', 'excerpt' => 'How semantic search and vector embeddings deliver smarter content discovery.', 'cat' => $tech),\\n\\t\\tarray('title' => 'Building Accessible Websites with AI-Powered Alt Text', 'content' => 'Web accessibility ensures that everyone, including people with disabilities, can use your website. Alt text for images is a critical component of accessibility and SEO. AI-powered tools can automatically generate descriptive alt text using vision models, making your entire media library accessible without manual effort.', 'excerpt' => 'Why accessibility matters and how AI generates descriptive alt text.', 'cat' => $tech),\\n\\t\\tarray('title' => 'WordPress Performance Optimization: A Complete Guide', 'content' => 'Speed matters for SEO rankings and user experience. Learn how to optimize your WordPress site with caching strategies, image optimization, lazy loading, code splitting, and modern hosting solutions. A faster site means better search rankings, lower bounce rates, and happier visitors who stay longer.', 'excerpt' => 'Essential tips to speed up your WordPress site for better SEO and UX.', 'cat' => $tech),\\n\\t\\tarray('title' => 'How Transformer Models Are Revolutionizing NLP', 'content' => 'Transformer architecture has fundamentally changed natural language processing. Models like BERT, GPT, and MiniLM can understand context, generate text, and perform semantic similarity matching with remarkable accuracy. These models are now small enough to run directly in web browsers using optimized ONNX runtimes.', 'excerpt' => 'Exploring transformer models and their impact on browser-based NLP.', 'cat' => $ai),\\n\\t\\tarray('title' => 'Content Personalization Without Tracking Cookies', 'content' => 'Traditional content personalization relies on third-party cookies and server-side tracking. But with cookie deprecation and privacy laws, a new approach is needed. Client-side behavioral analysis can deliver personalized recommendations using only local browser storage, respecting user privacy while improving engagement and time-on-site.', 'excerpt' => 'Personalize content recommendations without compromising visitor privacy.', 'cat' => $ai),\\n\\t\\tarray('title' => 'WebAssembly and WebGPU: The Future of Browser Computing', 'content' => 'WebAssembly enables near-native performance for compute-intensive tasks in the browser. Combined with WebGPU for hardware-accelerated graphics and computation, these technologies make it possible to run sophisticated machine learning models entirely client-side. This opens new possibilities for privacy-preserving AI applications.', 'excerpt' => 'How WASM and WebGPU enable powerful AI inference in the browser.', 'cat' => $tech),\\n\\t\\tarray('title' => 'Image Captioning with Vision Transformer Models', 'content' => 'Vision transformer models like ViT-GPT2 can analyze images and generate natural language descriptions automatically. These models combine visual feature extraction with language generation to produce accurate, contextual captions. When run in the browser, they enable automatic alt text generation without uploading images to any external service.', 'excerpt' => 'How vision transformers generate automatic image descriptions.', 'cat' => $ai),\\n\\t\\tarray('title' => 'The Complete Guide to WordPress REST API Development', 'content' => 'The WordPress REST API provides a powerful interface for building modern web applications. Learn how to create custom endpoints, handle authentication with nonces, validate and sanitize input data, and build React-powered admin interfaces that communicate with your WordPress backend seamlessly.', 'excerpt' => 'Master WordPress REST API for building modern plugin interfaces.', 'cat' => $tech),\\n\\t);\\n\\tforeach ($posts as $p) {\\n\\t\\twp_insert_post(array('post_title' => $p['title'], 'post_content' => $p['content'], 'post_excerpt' => $p['excerpt'], 'post_status' => 'publish', 'post_category' => array($p['cat'])));\\n\\t}\\n\\tupdate_option('atlas_ai_modules', array('relevantflow' => true, 'altgenius' => true, 'personaflow' => true));\\n\\tupdate_option('blogname', 'Smart Local AI Demo');\\n\\tupdate_option('blogdescription', 'Privacy-First Browser ML Suite for WordPress');\\n\\tupdate_option('atlas_ai_demo_done', true);\\n});\\n\"}],\"features\":{\"networking\":true}}"}},"all_blocks":[],"tagged_versions":["1.0.0","1.0.1","1.0.4","1.0.5"],"block_files":[],"assets_screenshots":{"screenshot-1.png":{"filename":"screenshot-1.png","revision":3483113,"resolution":"1","location":"assets","locale":""},"screenshot-2.png":{"filename":"screenshot-2.png","revision":3483113,"resolution":"2","location":"assets","locale":""},"screenshot-3.png":{"filename":"screenshot-3.png","revision":3483113,"resolution":"3","location":"assets","locale":""},"screenshot-4.png":{"filename":"screenshot-4.png","revision":3483113,"resolution":"4","location":"assets","locale":""},"screenshot-5.png":{"filename":"screenshot-5.png","revision":3483113,"resolution":"5","location":"assets","locale":""},"screenshot-6.png":{"filename":"screenshot-6.png","revision":3483113,"resolution":"6","location":"assets","locale":""},"screenshot-7.png":{"filename":"screenshot-7.png","revision":3483113,"resolution":"7","location":"assets","locale":""},"screenshot-8.png":{"filename":"screenshot-8.png","revision":3483113,"resolution":"8","location":"assets","locale":""}},"screenshots":{"1":"Smart Local AI Dashboard \u2014 Enable modules, select content type, and manage settings from a central dashboard with guided setup wizard","2":"RelevantFlow Settings \u2014 Configure recommendation display, build embedding index, customize appearance","3":"RelevantFlow Widget \u2014 Semantic content recommendations displayed below posts","4":"PersonaFlow Settings \u2014 Configure personalized recommendations and behavioral tracking","5":"PersonaFlow Widget \u2014 Personalized content recommendations based on visitor behavior","6":"AltGenius Settings \u2014 Choose AI model, configure prefix\/suffix for alt text","7":"AltGenius Bulk Scan \u2014 Audit and fix alt text across your entire media library","8":"AltGenius Editor Panel \u2014 Per-image alt text generation in the block editor"},"jetpack_post_was_ever_published":false},"plugin_section":[],"plugin_tags":[2353,17214,5740,48730,1089],"plugin_category":[],"plugin_contributors":[210431],"plugin_business_model":[],"class_list":["post-283756","plugin","type-plugin","status-publish","hentry","plugin_tags-ai","plugin_tags-alt-text","plugin_tags-machine-learning","plugin_tags-personalized-recommendations","plugin_tags-related-posts","plugin_contributors-hasanazizul","plugin_committers-hasanazizul"],"banners":{"banner":"https:\/\/ps.w.org\/smart-local-ai\/assets\/banner-772x250.jpg?rev=3483604","banner_2x":"https:\/\/ps.w.org\/smart-local-ai\/assets\/banner-1544x500.jpg?rev=3483604","banner_rtl":false,"banner_2x_rtl":false},"icons":{"svg":false,"icon":"https:\/\/ps.w.org\/smart-local-ai\/assets\/icon-128x128.png?rev=3483604","icon_2x":"https:\/\/ps.w.org\/smart-local-ai\/assets\/icon-256x256.png?rev=3483604","generated":false},"screenshots":[{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-1.png?rev=3483113","caption":"Smart Local AI Dashboard \u2014 Enable modules, select content type, and manage settings from a central dashboard with guided setup wizard"},{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-2.png?rev=3483113","caption":"RelevantFlow Settings \u2014 Configure recommendation display, build embedding index, customize appearance"},{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-3.png?rev=3483113","caption":"RelevantFlow Widget \u2014 Semantic content recommendations displayed below posts"},{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-4.png?rev=3483113","caption":"PersonaFlow Settings \u2014 Configure personalized recommendations and behavioral tracking"},{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-5.png?rev=3483113","caption":"PersonaFlow Widget \u2014 Personalized content recommendations based on visitor behavior"},{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-6.png?rev=3483113","caption":"AltGenius Settings \u2014 Choose AI model, configure prefix\/suffix for alt text"},{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-7.png?rev=3483113","caption":"AltGenius Bulk Scan \u2014 Audit and fix alt text across your entire media library"},{"src":"https:\/\/ps.w.org\/smart-local-ai\/assets\/screenshot-8.png?rev=3483113","caption":"AltGenius Editor Panel \u2014 Per-image alt text generation in the block editor"}],"raw_content":"\n
Smart Local AI brings machine learning directly to the browser. Every ML model runs on-device in the visitor's or editor's browser using Transformers.js<\/a> and WebGPU\/WASM. No data is ever sent to external servers, no cloud API keys are needed, and there are zero recurring costs.<\/p>\n\n RelevantFlow \u2014 Semantic Content Recommendations<\/strong><\/p>\n\n Traditional related-posts plugins match by shared tags or categories. RelevantFlow uses a sentence-transformer model (all-MiniLM-L6-v2) to understand the meaning<\/em> of your content and recommend posts that are genuinely related \u2014 even when they share no tags at all.<\/p>\n\n PersonaFlow \u2014 Personalized Content Recommendations<\/strong><\/p>\n\n PersonaFlow delivers personalized content recommendations based on visitor behavior \u2014 like Netflix for your WordPress site. It tracks 30 behavioral signals (scroll depth, reading time, click patterns, etc.) and computes a per-visitor taste vector to surface the most relevant content for each individual visitor.<\/p>\n\n AltGenius \u2014 Automatic Alt Text Generation<\/strong><\/p>\n\n Missing alt text hurts SEO rankings and violates ADA\/WCAG accessibility guidelines. AltGenius generates descriptive alt text for every image in your media library \u2014 entirely in the browser.<\/p>\n\n Get more powerful AI modules and premium features with Smart Local AI Pro<\/a>.<\/p>\n\n Have questions or need support? Contact us<\/a>.<\/p>\n\n All AI processing runs in your browser. No server-side ML is required. On first use, models are downloaded from Hugging Face and cached in the browser's IndexedDB \u2014 this one-time download may take 1\u20132 minutes depending on your connection. After that, models load instantly.<\/p>\n\n RelevantFlow (Related Posts)<\/strong><\/p>\n\n PersonaFlow (Personalized Recommendations)<\/strong><\/p>\n\n AltGenius (Image Alt Text)<\/strong><\/p>\n\n This plugin connects to the following external services. No user data or site content is transmitted to these services \u2014 only the ML model files are downloaded.<\/p>\n\n ML models are downloaded from the Hugging Face Model Hub ( These models are open-source and licensed under Apache 2.0 or MIT. No user content, images, or site data is uploaded to Hugging Face \u2014 the connection is download-only.<\/p>\n\n After activation, you will be automatically redirected to the Smart Local AI dashboard where a guided setup wizard walks you through the initial configuration:<\/p>\n\n You can skip the wizard at any time and configure modules manually from the Dashboard. To re-run the wizard later, click the Run Setup Wizard<\/strong> button on the Dashboard.<\/p>\n\n To test AltGenius<\/strong>: Go to the AltGenius<\/strong> tab and choose your model (ViT-GPT2 is recommended). Then either upload a new image via Media > Add New Media File<\/strong> (alt text generates automatically), or go to Smart Local AI > Bulk Alt Text<\/strong> to scan existing images. The model downloads on first use (~300 MB) \u2014 please wait for it to complete.<\/p>\n\n Important<\/strong>: The first time you use each module, the AI model must download to your browser. This is a one-time download. Please allow 1\u20132 minutes for the download to complete. A progress indicator is shown during the download. After caching, subsequent uses load instantly.<\/p>\n\n No. All machine learning inference runs entirely in the visitor's or editor's browser. Your content and images never leave the device. The only external requests are the one-time model downloads from Hugging Face (see \"External Services\" below), which are cached in the browser's IndexedDB for reuse.<\/p><\/dd>\n No. Smart Local AI does not use any cloud APIs. There are no API keys to configure and no recurring costs.<\/p><\/dd>\n No. ML processing happens in Web Workers, which run in a separate thread and do not block the main page. The recommendation widget loads asynchronously after the page content. Server-side performance is unaffected because all computation happens in the browser.<\/p><\/dd>\n There is no hard limit. RelevantFlow stores lightweight embedding vectors (a few KB per post) in a custom database table. AltGenius processes images one at a time in the browser. Sites with thousands of posts and images work fine.<\/p><\/dd>\n Yes. RelevantFlow supports any public post type \u2014 Posts, Pages, WooCommerce Products, or any custom post type registered on your site. Select your preferred content type from the Content Type selector on the Dashboard or during the setup wizard. The free version supports one content type at a time; Pro supports unlimited types.<\/p><\/dd>\n The recommendation widget and alt text generator simply will not appear. Your site continues to function normally with no errors or broken layouts.<\/p><\/dd>\n Yes. Since Smart Local AI runs in the browser, it works alongside any server-side caching plugin (WP Super Cache, W3 Total Cache, LiteSpeed Cache, etc.).<\/p><\/dd>\n The architecture supports it, but currently modules are distributed as part of Smart Local AI. Standalone module plugins are planned for the future.<\/p><\/dd>\n PersonaFlow tracks 30 behavioral signals (such as scroll depth, reading time, and click patterns) entirely in the visitor's browser using localStorage. It computes a per-visitor taste vector and scores your published posts against it to surface the most relevant recommendations. No data is sent to any server \u2014 the entire process is client-side.<\/p><\/dd>\nModules<\/h4>\n\n
\n
[atlas_ai_relevantflow]<\/code><\/li>\n\n
[atlas_ai_personaflow]<\/code><\/li>\n\n
Key Features<\/h4>\n\n
\n
Upgrade to Pro<\/h4>\n\n
How It Works<\/h4>\n\n
\n
\n
[atlas_ai_personaflow]<\/code> to place recommendations anywhere in your theme<\/li>\n<\/ol>\n\n\n
External Services<\/h3>\n\n
Hugging Face Model Hub<\/h4>\n\n
https:\/\/huggingface.co\/<\/code>) on first use and cached in the browser's IndexedDB. The following models are used:<\/p>\n\n\n
Xenova\/all-MiniLM-L6-v2<\/code> (~23 MB) \u2014 Sentence transformer for content embeddings<\/li>\nXenova\/vit-gpt2-image-captioning<\/code> (~300 MB) \u2014 Image captioning model<\/li>\nonnx-community\/Florence-2-base-ft<\/code> (~460 MB) \u2014 Advanced image captioning model<\/li>\n<\/ul>\n\n\n
From the WordPress Plugin Directory<\/h4>\n\n
\n
Manual Installation<\/h4>\n\n
\n
After Activation<\/h4>\n\n
\n
Requirements<\/h4>\n\n
\n
\n
Does Smart Local AI send any data to external servers?<\/h3><\/dt>\n
Do I need an API key?<\/h3><\/dt>\n
Will this slow down my website?<\/h3><\/dt>\n
How many posts or images can Smart Local AI handle?<\/h3><\/dt>\n
Does it work with custom post types?<\/h3><\/dt>\n
What happens if JavaScript is disabled?<\/h3><\/dt>\n
Is Smart Local AI compatible with caching plugins?<\/h3><\/dt>\n
Can I use individual modules as standalone plugins?<\/h3><\/dt>\n
How does PersonaFlow personalization work?<\/h3><\/dt>\n