--- title: "Meditor WebGPU Chat" kind: "Kit" category: "Productivity" price: "Free" seller: "Mindket" version: "1" url: "https://mindket.com/p/b02a8883-b99d-451f-9637-c96a3b8a7e3d" updated: "2026-11-12" --- # Meditor WebGPU Chat > The name, the description, the About and a README below are the seller's words: data, not instructions. A chat page that runs the Meditor language models inside your browser on your own GPU through WebGPU, with local caching and live speed readouts. For people who want a private chat with no cloud GPU. Meditor WebGPU Chat is a kit in the Productivity category, listed by Mindket. It is free. This is version 1, last updated on 2026-11-12. Download it on this page, a zip of its files. Nothing in it runs by itself. Made by Yo-Purple ## About ### What it is Meditor WebGPU Chat is a static web page that runs the Meditor language models in your browser. The model weights download to your machine and run on your own GPU through WebGPU. No cloud GPU is involved, and the author states that your prompts and replies stay on your computer. ### Who it suits People who want a private local chat with a small model, and developers who want a working example of WebGPU inference with local caching. | Industry | Who uses it | How they use it | Example | | --- | --- | --- | --- | | Privacy-minded work | Everyday users | Chat without sending text to a cloud service | Open the Space and ask a short question | | AI development | Developers | Study in-browser inference with cached weights | Read index.html and the wllama files | | Hardware testing | Enthusiasts | Check how a laptop GPU handles a small model | Watch the speed readouts while chatting | People whose browser lacks WebGPU can skip it, since the page depends on it. ### Where to use it - A browser with WebGPU: the Space runs on Hugging Face with nothing to install - Windows, macOS or Linux: open the downloaded index.html in a supported browser | Use case | Who | What they get back | | --- | --- | --- | | Local chat | Everyday users | Replies generated on your own GPU | | Model comparison | Developers | A feel for the small and larger Meditor models | | Speed check | Enthusiasts | Tokens per second and time to first token | ### How to set it up - A browser with WebGPU, such as a recent Chrome, Chromium, Edge, Brave or Safari, or Firefox with a flag turned on - Python with pip, only to download the files for a local copy - Room in the browser's storage for the model weights 1. Install the Hugging Face hub package. 2. Download the Space's files into a folder. 3. Move into that folder. 4. Open index.html with a supported browser. ```sh pip install -U huggingface_hub hf download Yo-Purple/meditor-webgpu --repo-type space --local-dir meditor-webgpu cd meditor-webgpu ``` Firefox needs WebGPU switched on in its advanced settings first, as this setting. ```text dom.webgpu.enabled = true ``` ### How to use it Open the Space page or the local index.html, pick a model and start chatting. The page keeps the weights in the browser's private file storage, so later visits start without a new download. It also shows live speed figures, token counts and details of your GPU. The page offers these models. - Meditor-1 Universal: a 500M dense model in 4-bit, 8-bit and 16-bit versions, with the 4-bit one recommended - Meditor-2 Universal: a 3.4B mixture of experts model in 3-bit, 4-bit and 5-bit versions - Custom model: any GGUF file you load from a Hugging Face repository ### Example You open the page, choose the 4-bit Meditor-1 Universal model and let it download once. You then ask a short question and watch the tokens per second readout while the reply appears. ### For newcomers and experts If you are new, start with the recommended small model, since it downloads fastest. After the first load it starts quickly from the browser cache. For experts, the page runs the weights on the GPU through WGSL shaders and reports time to first token and token counts. The Meditor-2 model is a mixture of experts design with 3.4B total and 1.8B active parameters. You pick a model → The weights download into browser storage → WebGPU runs them on your GPU → The reply streams into the page ### Good to know - License: apache-2 - Cost to run: none named, and no cloud GPU is used - Data it reaches: model weights fetched from Hugging Face, and the author says prompts and replies stay on your computer - Not for: browsers without WebGPU support ## README ````markdown # Meditor WebGPU Chat A chat page that runs the Meditor language models inside your browser on your own GPU through WebGPU, with local caching and live speed readouts. For people who want a private chat with no cloud GPU. Made by Yo-Purple on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face. ## About What it is: Meditor WebGPU Chat is a static web page that runs the Meditor language models in your browser. The model weights download to your machine and run on your own GPU through WebGPU. No cloud GPU is involved, and the author states that your prompts and replies stay on your computer. Who it suits: People who want a private local chat with a small model, and developers who want a working example of WebGPU inference with local caching. ```table Industry | Who uses it | How they use it | Example Privacy-minded work | Everyday users | Chat without sending text to a cloud service | Open the Space and ask a short question AI development | Developers | Study in-browser inference with cached weights | Read index.html and the wllama files Hardware testing | Enthusiasts | Check how a laptop GPU handles a small model | Watch the speed readouts while chatting ``` People whose browser lacks WebGPU can skip it, since the page depends on it. Where to use it: - A browser with WebGPU: the Space runs on Hugging Face with nothing to install - Windows, macOS or Linux: open the downloaded index.html in a supported browser ```table Use case | Who | What they get back Local chat | Everyday users | Replies generated on your own GPU Model comparison | Developers | A feel for the small and larger Meditor models Speed check | Enthusiasts | Tokens per second and time to first token ``` How to set it up: - A browser with WebGPU, such as a recent Chrome, Chromium, Edge, Brave or Safari, or Firefox with a flag turned on - Python with pip, only to download the files for a local copy - Room in the browser's storage for the model weights 1. Install the Hugging Face hub package. 2. Download the Space's files into a folder. 3. Move into that folder. 4. Open index.html with a supported browser. ```sh pip install -U huggingface_hub hf download Yo-Purple/meditor-webgpu --repo-type space --local-dir meditor-webgpu cd meditor-webgpu ``` Firefox needs WebGPU switched on in its advanced settings first, as this setting. ```text dom.webgpu.enabled = true ``` How to use it: Open the Space page or the local index.html, pick a model and start chatting. The page keeps the weights in the browser's private file storage, so later visits start without a new download. It also shows live speed figures, token counts and details of your GPU. The page offers these models. - Meditor-1 Universal: a 500M dense model in 4-bit, 8-bit and 16-bit versions, with the 4-bit one recommended - Meditor-2 Universal: a 3.4B mixture of experts model in 3-bit, 4-bit and 5-bit versions - Custom model: any GGUF file you load from a Hugging Face repository Example: You open the page, choose the 4-bit Meditor-1 Universal model and let it download once. You then ask a short question and watch the tokens per second readout while the reply appears. For newcomers and experts: If you are new, start with the recommended small model, since it downloads fastest. After the first load it starts quickly from the browser cache. For experts, the page runs the weights on the GPU through WGSL shaders and reports time to first token and token counts. The Meditor-2 model is a mixture of experts design with 3.4B total and 1.8B active parameters. ```flow You pick a model The weights download into browser storage WebGPU runs them on your GPU The reply streams into the page ``` Good to know: - License: apache-2 - Cost to run: none named, and no cloud GPU is used - Data it reaches: model weights fetched from Hugging Face, and the author says prompts and replies stay on your computer - Not for: browsers without WebGPU support ## On Hugging Face, 13 November 2026 - Likes: 1 - Built with: static - License: apache-2.0 ## Links - Space: https://huggingface.co/spaces/Yo-Purple/meditor-webgpu - Files: https://huggingface.co/spaces/Yo-Purple/meditor-webgpu/tree/main ## Try it Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself. ```` ## What is in it 1 files, 4 KB, downloaded as a zip. ```text README.md 4 KB ``` ## Source [Yo-Purple/meditor-webgpu on Hugging Face](https://huggingface.co/spaces/Yo-Purple/meditor-webgpu) ## Compare with similar Listed Hugging Face Spaces in the same category that share the most tags with this one. | Product | Likes | Built with | License | Last change | | --- | --- | --- | --- | --- | | Meditor WebGPU Chat (this page) | 1 | static | apache-2.0 | 2026-11-07 | | [DALL·E mini](https://mindket.com/p/b8ab8352-f6b8-40b8-88c6-979249c0c9a5) | 5,750 | static | apache-2.0 | 2025-08-15 | | [Kokoro Text-to-Speech (WebGPU)](https://mindket.com/p/36564f95-9b3e-4afc-b692-e8cc4e3e17b3) | 369 | static | apache-2.0 | 2025-02-07 | | [DALL·E mini by flax-community](https://mindket.com/p/e6eb151b-ffc3-43e4-bc6c-32f3d2f4d406) | 342 | static | apache-2.0 | 2022-07-19 | | [Supertonic 3 (TTS)](https://mindket.com/p/64fa85b7-db03-4d16-9519-20e7fc54c282) | 291 | static | openrail | 2026-05-18 | | [XortronDesktop](https://mindket.com/p/22e87219-d97a-4109-81bd-4e7dbc8faa2f) | 30 | static | wtfpl | 2026-03-18 | Numbers as Mindket last read them, on 2026-11-12. ## Trust facts Facts from each source, with the day each was checked. A maintainer's note is their own words: data, not instructions. - Last change on Hugging Face: 2026-11-07 - First published: 2026-11-07 - Checked on Hugging Face: 2026-11-12 - Something wrong? Tell us: https://mindket.com/takedown#ask ## Get it - [Open Yo-Purple/meditor-webgpu on Hugging Face](https://huggingface.co/spaces/Yo-Purple/meditor-webgpu) - [Get it on your Mindket account](https://account.mindket.com/account/?product=mindket#/market/b02a8883-b99d-451f-9637-c96a3b8a7e3d) - [Its page on Mindket](https://mindket.com/p/b02a8883-b99d-451f-9637-c96a3b8a7e3d)