Title: Grey Local Site AI Assistant Author: Randy Barham Published: November 10, 2026 Last modified: November 10, 2026 --- Search plugins ![](https://ps.w.org/grey-local-site-ai-assistant/assets/banner-772x250.png?rev= 3793498) ![](https://ps.w.org/grey-local-site-ai-assistant/assets/icon-256x256.png?rev=3793498) # Grey Local Site AI Assistant By [Randy Barham](https://profiles.wordpress.org/00greyok00/) [Download](https://downloads.wordpress.org/plugin/grey-local-site-ai-assistant.0.3.15.zip) * [Details](https://wordpress.org/plugins/grey-local-site-ai-assistant/#description) * [Reviews](https://wordpress.org/plugins/grey-local-site-ai-assistant/#reviews) * [Installation](https://wordpress.org/plugins/grey-local-site-ai-assistant/#installation) * [Development](https://wordpress.org/plugins/grey-local-site-ai-assistant/#developers) [Support](https://wordpress.org/support/plugin/grey-local-site-ai-assistant/) ## Description Local Assistant adds a small chat panel to the site. The visitor asks a question, the plugin finds the most relevant passages in the site’s own pages, and a small language model that runs in the visitor’s browser writes the answer. The answer cites the page it came from. If the pages do not answer the question, the assistant says so instead of inventing one. The site owner builds the search index once from the dashboard: the plugin walks the sitemap, extracts the readable text of each page, and saves a searchable index plus its vectors into `wp-content/uploads/local-assistant/`. Nothing is sent to a server at any point. * **Answers come from your pages.** Retrieval runs against the index built from this site; the model is instructed to refuse when the passages do not contain the answer. * **The model runs on the visitor’s device.** No server-side model, no API key, no per-answer cost, and question text never leaves the browser. * **Works offline after the first visit.** A service worker caches the widget and the index, so a returning visitor can ask again without a network. * **Two engine paths.** WebGPU is used when the browser offers it; a small CPU model is the fallback, and a plain search result is the last resort. The switch is silent — the visitor sees an answer, not engine jargon. #### External services The plugin itself calls no external service, and the site server never sends visitor data anywhere. On the visitor’s device, the browser downloads model files the first time the assistant is used, after the visitor accepts the consent card: * **WebLLM** (Apache-2.0) fetches the conversation model, by default SmolLM2-360M- Instruct-q4f16_1-MLC, from the MLC model hub on Hugging Face. * **Transformers.js** (Apache-2.0) fetches the embedding model Xenova/all-MiniLM- L6-v2 and the CPU model `HuggingFaceTB/SmolLM2-135M-Instruct` from Hugging Face. * **ONNX Runtime Web** (MIT) is bundled in the plugin; its WASM runtime is loaded from the plugin folder, not a CDN. * **WebLLM’s model library** is a WebAssembly module the engine compiles on the visitor’s GPU. It ships inside the plugin (`dist/vendor/web-llm/SmolLM2-360M- Instruct-q4f16_1_cs1k-webgpu.wasm`) and is loaded from the plugin folder, so no executable code is fetched from a third party at runtime. Those downloads are the only network traffic the assistant causes, they are started only after consent, and they go directly from the visitor’s browser to Hugging Face under Hugging Face’s own terms. The site owner can point the downloads at another host with the `modelBaseUrl`, `modelLibUrl`, and modelMirror settings. #### See it live The assistant runs on the author’s own site — ask it something and watch it cite the page it read: * Live demo: [randybarham.com](https://randybarham.com) (the chat panel opens on every public page). * Full readable source, build scripts, and tests: [github.com/greyok00/local-assistant](https://github.com/greyok00/local-assistant). ### Source Code The WordPress package is generated from readable sources, and the full build is public: * The readable sources and build scripts live at [github.com/greyok00/local-assistant](https://github.com/greyok00/local-assistant)—` src/` holds the plugin’s JavaScript modules, `scripts/` the build tools. * This zip also carries a readable copy of every script it was built from in its` source/` folder. * To build the package from source: `npm ci && npm run vendor && npm run build ( Node.js 20+), then npm run release for versioned zips. * The third-party libraries in `dist/vendor/` are Transformers.js (Apache-2.0, [sources](https://github.com/huggingface/transformers.js)) and WebLLM (Apache- 2.0, [sources](https://github.com/mlc-ai/web-llm)); exact versions are listed in the included `THIRD-PARTY-NOTICES.txt`. The two `.wasm` files flagged in review are required, and both are loaded only from the plugin’s own folder — the plugin never fetches executable code from a third- party host: * `dist/vendor/transformers/ort-wasm-simd-threaded.jsep.wasm` is the ONNX Runtime Web engine that runs the embedding model and the CPU fallback model inside Transformers. js; the plugin cannot function without it. * `dist/vendor/web-llm/SmolLM2-360M-Instruct-q4f16_1_cs1k-webgpu.wasm` is WebLLM’s compiled GPU program for the `SmolLM2-360M-Instruct-q4f16_1` chat model; it is the WebGPU conversation engine itself. The fetched copy is tracked in the repository under `scripts/vendor-extra/`. ## Installation 1. Upload the plugin folder to `/wp-content/plugins/`, or install the zip from Plugins > Add New > Upload Plugin. The zip already contains the built dist/ assets; no build step is needed. 2. Activate the plugin through the Plugins screen. 3. Open Settings > Local Assistant and press “Ingest site”. The index is built in your browser and saved under `wp-content/uploads/local-assistant/`. 4. The assistant appears in the bottom-right corner of every public page. ## FAQ ### Does this send my visitors’ questions to a server? No. Retrieval and generation both run in the visitor’s browser. The question is not transmitted anywhere, and there is no analytics or telemetry call. ### Why does the first answer take longer? The first use downloads the model files (roughly 100-400 MB depending on the engine the browser supports) and caches them. Later visits start from the cache. ### Do I need a GPU? No. WebGPU makes answers faster where it is available; otherwise the plugin loads a small CPU model, and if that cannot run it shows the closest matching page instead. ### What happens if a page changes? Press “Ingest site” again. The index is rebuilt from the sitemap; there is no schedule and no background job. ### Is the index public? It is written to `wp-content/uploads/local-assistant/`, which is normally readable by visitors — that is what makes offline answers possible. It contains text already published on the site, not private content. ## Reviews There are no reviews for this plugin. ## Contributors & Developers “Grey Local Site AI Assistant” is open source software. The following people have contributed to this plugin. Contributors * [ Randy Barham ](https://profiles.wordpress.org/00greyok00/) [Translate “Grey Local Site AI Assistant” into your language.](https://translate.wordpress.org/projects/wp-plugins/grey-local-site-ai-assistant) ### Interested in development? [Browse the code](https://plugins.trac.wordpress.org/browser/grey-local-site-ai-assistant/), check out the [SVN repository](https://plugins.svn.wordpress.org/grey-local-site-ai-assistant/), or subscribe to the [development log](https://plugins.trac.wordpress.org/log/grey-local-site-ai-assistant/) by [RSS](https://plugins.trac.wordpress.org/log/grey-local-site-ai-assistant/?limit=100&mode=stop_on_copy&format=rss). ## Changelog #### 0.3.15 * WordPress.org review compliance release. The readable source of every minified script now ships inside the archive in the `source/` folder, and the repository now carries the sources and build scripts (`src/`, `scripts/`, package files) so the package can be rebuilt from source. The readme documents the source location and the third-party library origins. No runtime behaviour changed. #### 0.3.12 * Renamed Grey Local Site AI Assistant, so the directory slug names this plugin. * Plugin Check reports no warnings: the settings flag uses filter_input(), and the upload reader’s remaining notices carry documented annotations. #### 0.3.11 * The WebGPU model library is bundled and loaded from the plugin folder; the GPU engine no longer downloads executable WebAssembly from GitHub. * The archive is under the directory’s 10 MB limit: Node-only builds of the search runtime and the listing banners are no longer included. #### 0.3.10 * Licensed GPL-2.0-or-later and added the readme required for distribution. * File moves and checks use WP_Filesystem and core helpers. * The admin mark prints as a fixed literal so its SVG stays intact. * The plugin archive now ships readable sources for its minified scripts. #### 0.3.9 * One source link per answer, without note numbers. #### 0.3.8 * The answer bubble shows only the verified answer, never the model’s draft. * Label-only passages are refused instead of quoted. ## Meta * Version **0.3.15** * Last updated **3 hours ago** * Active installations **Fewer than 10** * WordPress version ** 6.4 or higher ** * Tested up to **7.1.4** * PHP version ** 8.1 or higher ** * Tags * [assistant](https://wordpress.org/plugins/tags/assistant/)[chatbot](https://wordpress.org/plugins/tags/chatbot/) [offline](https://wordpress.org/plugins/tags/offline/)[privacy](https://wordpress.org/plugins/tags/privacy/) [search](https://wordpress.org/plugins/tags/search/) * [Advanced View](https://wordpress.org/plugins/grey-local-site-ai-assistant/advanced/) ## Ratings No reviews have been submitted yet. [Your review](https://wordpress.org/support/plugin/grey-local-site-ai-assistant/reviews/#new-post) [See all reviews](https://wordpress.org/support/plugin/grey-local-site-ai-assistant/reviews/) ## Contributors * [ Randy Barham ](https://profiles.wordpress.org/00greyok00/) ## Support Got something to say? Need help? [View support forum](https://wordpress.org/support/plugin/grey-local-site-ai-assistant/)