Loading video...

Video Failed to Load

Go Home

We just open-sourced the world's fastest WebGPU kernels for 200+ machine learning operations. Specialized for your hardware. They run entirely locally in your browser. Now available on Hugging Face 🤗 Tell your agent to use @huggingface/kernels in your next project.

xenovacom's profile picture

Xenova

20,786 subscribers

20,104 views • 2 days ago •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

7 Comments

Xenova's profile picture
Xenova2 days ago

We're also working to upstream these optimizations to Transformers.js, ONNX Runtime Web, LiteRT.js, and more! Kernels: Blog:

Julien Blanchon 🇺🇦's profile picture
Julien Blanchon 🇺🇦2 days ago

I guess this is forward only, no backward ?

Zack Fitch's profile picture
Zack Fitch2 days ago

WebGPU is life

Jens Fursund's profile picture
Jens Fursund2 days ago

Oooh super nice!

soulblocks's profile picture
soulblocks2 days ago

Oh damn 🔥

banzaan's profile picture
banzaan2 days ago

browser is turning into an entire os running local models there changes everything wild drop

Harsha's profile picture
Harsha2 days ago

local kernels make the browser a more interesting compute boundary. i'd love a demo where the model runs on private input locally and the agent gets only the result it needs. how much of the workflow can stay inside that boundary?

Related Videos

It's official: Qwen3.8-27B just scored 52 on the Artificial Analysis Intelligence Index. We now have an open-weight model that matches GPT-5.6 Luna (max) AND can run locally... even in your browser with custom WebGPU kernels! What a time to be alive! 🤯
0:58

It's official: Qwen3.8-27B just scored 52 on the Artificial Analysis Intelligence Index. We now have an open-weight model that matches GPT-5.6 Luna (max) AND can run locally... even in your browser with custom WebGPU kernels! What a time to be alive! 🤯

Xenova

187,874 views • 1 month ago

made a site to run deepseek r1 for free -- locally in your browser, no downloads, no servers purely using WebGPU r1-web is open source, made in america, run on american servers, available below forever
1:21

made a site to run deepseek r1 for free -- locally in your browser, no downloads, no servers purely using WebGPU r1-web is open source, made in america, run on american servers, available below forever

Surya

440,240 views • 1 year ago

Introducing: a coding agent (Pi) running entirely in your browser using a 2B model on WebGPU 🤯 MiniCPM5-2B + Pi, powered by Transformers.js + WebGPU + 4-bit ONNX weights. All previous attempt to create this failed but MiniCPM5 seems to make it usable. Available now on Hugging Face 👇
0:23

Introducing: a coding agent (Pi) running entirely in your browser using a 2B model on WebGPU 🤯 MiniCPM5-2B + Pi, powered by Transformers.js + WebGPU + 4-bit ONNX weights. All previous attempt to create this failed but MiniCPM5 seems to make it usable. Available now on Hugging Face 👇

Victor M

89,700 views • 18 days ago

Before Fable 5 was shut down, it pushed Gemma 4 to 255 tok/s on WebGPU. Some didn't believe it was real. Today we're releasing the demo and kernels it wrote for you to see yourself. Run it locally in your browser. Agentic kernel optimization is the future of on-device inference
0:57

Before Fable 5 was shut down, it pushed Gemma 4 to 255 tok/s on WebGPU. Some didn't believe it was real. Today we're releasing the demo and kernels it wrote for you to see yourself. Run it locally in your browser. Agentic kernel optimization is the future of on-device inference

Xenova

485,900 views • 3 months ago

Introducing the new dev-browser cli. The fastest way for an agent to use a browser is to let it write code. Just `npm i -g dev-browser` and tell your agent to "use dev-browser"
0:30

Introducing the new dev-browser cli. The fastest way for an agent to use a browser is to let it write code. Just `npm i -g dev-browser` and tell your agent to "use dev-browser"

Sawyer Hood

870,855 views • 6 months ago

Bonsai 27B just changed the local LLM game forever. 1-bit quantization shrinks it from 54GB to just 3.8GB (-93%), while retaining 90% of its intelligence. That's insane. With custom WebGPU kernels written by Fable 5 and GPT 5.6 Sol, the model now runs locally in your browser!
1:29

Bonsai 27B just changed the local LLM game forever. 1-bit quantization shrinks it from 54GB to just 3.8GB (-93%), while retaining 90% of its intelligence. That's insane. With custom WebGPU kernels written by Fable 5 and GPT 5.6 Sol, the model now runs locally in your browser!

Xenova

211,255 views • 2 months ago

I made a site that uses WebGPU to run Qwen3 .6b with thinking locally, directly in your browser, no installation or servers necessary and runs offline Available forever for free and open source:
1:06

I made a site that uses WebGPU to run Qwen3 .6b with thinking locally, directly in your browser, no installation or servers necessary and runs offline Available forever for free and open source:

Surya Dantuluri

167,723 views • 10 months ago

NEW: Ternary Bonsai 2 just landed on Hugging Face — a 27B reasoning model in ternary weights, built for the agentic era 🤯 9x smaller than FP16 while retaining 98.2% of the intelligence. At <6GB in size, it can even run 100% locally in your browser on WebGPU! Try it yourself 👇
1:03

NEW: Ternary Bonsai 2 just landed on Hugging Face — a 27B reasoning model in ternary weights, built for the agentic era 🤯 9x smaller than FP16 while retaining 98.2% of the intelligence. At <6GB in size, it can even run 100% locally in your browser on WebGPU! Try it yourself 👇

Xenova

17,717 views • 15 days ago

Atomic Chat is now on Hugging Face 🤗 We're officially a Local App on the world's biggest AI hub. Run 200,000+ open-weight models from Hugging Face directly on your device - private, local, and open source!
0:22

Atomic Chat is now on Hugging Face 🤗 We're officially a Local App on the world's biggest AI hub. Run 200,000+ open-weight models from Hugging Face directly on your device - private, local, and open source!

atomic.chat

29,420 views • 3 months ago

AI is now FREE!! Genspark AI Browser now offers the world's first on-device free AI. - Zero internet required, run models locally on your machine - Download 169 open-weight models (GPT-OSS, Gemma3, etc.) - Private & Fast. Forever free for everyone. 1/3
1:40

AI is now FREE!! Genspark AI Browser now offers the world's first on-device free AI. - Zero internet required, run models locally on your machine - Download 169 open-weight models (GPT-OSS, Gemma3, etc.) - Private & Fast. Forever free for everyone. 1/3

Eric Jing

547,945 views • 1 year ago

NEW: Alibaba just released Qwen 3.5 Small — a family of powerful multimodal models available in a range of sizes (0.8B, 2B, 4B, and 9B parameters). Perfect for on-device applications! They can even run 100% locally in your browser on WebGPU, powered by Transformers.js! 🤯
0:46

NEW: Alibaba just released Qwen 3.5 Small — a family of powerful multimodal models available in a range of sizes (0.8B, 2B, 4B, and 9B parameters). Perfect for on-device applications! They can even run 100% locally in your browser on WebGPU, powered by Transformers.js! 🤯

Xenova

102,592 views • 7 months ago

I seriously cannot believe this is a 0.6B LLM! 🤯 Qwen just released Qwen3, a series of hybrid reasoning models that allow you to control how much "thinking" the model does for a given task. They can even run locally in your browser on WebGPU with 🤗 Transformers.js!
0:24

I seriously cannot believe this is a 0.6B LLM! 🤯 Qwen just released Qwen3, a series of hybrid reasoning models that allow you to control how much "thinking" the model does for a given task. They can even run locally in your browser on WebGPU with 🤗 Transformers.js!

Xenova

74,954 views • 1 year ago

Open AI released Operator, an agent that can use the browser to perform and automate tasks for you! I have built an Open Source version of Operator using Browser Use, running locally on your computer. 100% Open Source
0:59

Open AI released Operator, an agent that can use the browser to perform and automate tasks for you! I have built an Open Source version of Operator using Browser Use, running locally on your computer. 100% Open Source

Sumanth

62,485 views • 1 year ago

This doesn't feel like it should be possible. Muse Glimmer 30B can now run locally in your browser on WebGPU... at the same speed as llama.cpp (~25 tok/s on my M4 Max)! Just open a website. No installation required.
1:22

This doesn't feel like it should be possible. Muse Glimmer 30B can now run locally in your browser on WebGPU... at the same speed as llama.cpp (~25 tok/s on my M4 Max)! Just open a website. No installation required.

Xenova

21,313 views • 1 month ago

Happycapy is now open to everyone 🚀 The agent-native computer, in your browser, and now on your phone. > Powered by Claude Code + MiniMax, including Opus 4.6 and Minimax M2.5 > Your own secure cloud sandbox > Run agent teams. Automate task Not just another OpenClaw. It’s all-in-one. No installs. Just run.
1:42

Happycapy is now open to everyone 🚀 The agent-native computer, in your browser, and now on your phone. > Powered by Claude Code + MiniMax, including Opus 4.6 and Minimax M2.5 > Your own secure cloud sandbox > Run agent teams. Automate task Not just another OpenClaw. It’s all-in-one. No installs. Just run.

Jarod Xu

1,106,366 views • 7 months ago

Private AI browser with the OpenClaw agent on free local models Run your agent on Qwen, Gemma, or Nemotron directly in the browser Open source. Private. Runs on your local device
0:36

Private AI browser with the OpenClaw agent on free local models Run your agent on Qwen, Gemma, or Nemotron directly in the browser Open source. Private. Runs on your local device

Sigma Browser

115,689 views • 5 months ago

We open-sourced our agent skills ecosystem, built specifically for complex end-to-end browser tasks. Type in any domain, and researches an approach to reliably complete your task, contributing it back to the knowledge base for other agents. Watch our agent fetch Amazon listings in real-time, driven entirely by a custom browser skill:
0:20

We open-sourced our agent skills ecosystem, built specifically for complex end-to-end browser tasks. Type in any domain, and researches an approach to reliably complete your task, contributing it back to the knowledge base for other agents. Watch our agent fetch Amazon listings in real-time, driven entirely by a custom browser skill:

Shrey Pandya

10,312 views • 4 months ago

Introducing Anijam AI — the first AI Animation Agent on your phone. Your story idea should not die before you open professional software. Just tell Anijam your idea. It does the rest. Available now on iOS & browser.
2:29

Introducing Anijam AI — the first AI Animation Agent on your phone. Your story idea should not die before you open professional software. Just tell Anijam your idea. It does the rest. Available now on iOS & browser.

Anijam AI

260,758 views • 5 months ago

Superwhisper just released S1-mini, a 0.6B LLM designed to turn raw speech-to-text transcripts into clean written text. It can even run 100% locally in your browser on WebGPU, thanks to 🤗 Transformers.js! Try it out yourself! 👇
0:57

Superwhisper just released S1-mini, a 0.6B LLM designed to turn raw speech-to-text transcripts into clean written text. It can even run 100% locally in your browser on WebGPU, thanks to 🤗 Transformers.js! Try it out yourself! 👇

Xenova

13,286 views • 1 month ago