Apertus Mini is now running entirely in your browser 🇨🇭 80+ tps for the 1.5B, 60+ tps for the 4B (on my M3). Fully client-side via Transformers.js + ONNX + WebGPU.
🤷 Nico Martin
@nico.dev
Building Transformers.js at @hf.co // #GoogleDevExpert for Web Technologies and AI/ML 💚 #WebAI https://nico.dev
@tomayac.com colaborated with us on implementing the Cross-Origin Storage API in Transformers.js. Read his blogpost to learn how to try out the API right now🤗
Very excited to share a guest post I wrote for the Hugging Face 🤗 blog! Experimenting with the proposed Cross-Origin Storage API in Transformers.js! 👉 huggingface.co/blog/cross-o... This proposed API has the potential of revolutionizing the Web—a little bit. Learn more: github.com/WICG/cross-o...!
Everything you always wanted to know about Transformers.js, in one video. I made a deep dive into how AI models run from JavaScript: tensors, ONNX, quantization, `pipeline()`, WebGPU/WASM, preprocessing, postprocessing, and what happens under the hood. Watch: www.youtube.com/watch?v=A_6U...
How Transformers.js Works: AI Models in JavaScript, Explained
YouTube video by Hugging Face
youtube.com
While @xenova.bsky.social and I were distracted by Google I/O, 🤗 Transformers.js quietly crossed 1'000'000 weekly downloads 🤯.
Pretty wild.. I am 10km above Greenland while Googles Gemma 4 26b on llama.cpp connected to PI is helping me edit my video at 50tps 🤯
On my way to San Francisco🤩. First time in California. First time in Silicon Valley. And I'm going for Google I/O. 🤯 Also I recently started a new instagram channel. So if you want the unfiltered version, follow along👇😊
✨ new video in the making.. ...everything you always wanted to know about 🤗 Transformers.js
My learning from Uphill Conf '26: being a speaker is so much easier than being an organizer 😅 After the last two years on stage, I switched sides. Nine months of prep, a lot of group chats, and one very humbling ropes course at the speaker event later, we pulled it off!
I really enjoyed this conversation with the @neciudan.dev! We covered a lot.. how I went from Ski instructor to ML engineer and how Transformers.js lets you run ML models directly in the browser to WebGPU, quantization, agents, and what frontend devs should actually focus on🤗
Really cool episode! Thank you @nico.dev !!! 🚀🚀🚀 AI at Scale with Nico Martin from Hugging Face | Transformers.js, Tokenizers, On-Device Inference youtu.be/SilGBAV4Lbw
Getting started with AI in the browser has never been easier 🤗 Today we're releasing the Transformers.js #AgentSkill, your shortcut to running ML models in the browser. It provides all the information your coding agent needs to add on-device AI to your web app.
TranslateGemma 4B by Google now runs 100% in your browser on WebGPU with Transformers.js v4. 55 languages. No server. No data leaks. Works offline. A 4B parameter translation powerhouse, right in your browser. Try the demo 👇
Exciting news: this year I'm joining the organizing team of @uphillconf for the first time! 🎉 For the past two years, I've had the opportunity to give a talk, and each time it was a fantastic experience. That's why I decided to actively participate in the planning this year😊
I've seen several posts lately that share some edge cases in AI conversations where the AI makes mistakes. Of course, this is to show how bad AI still is. AI is a tool. If it produces incorrect or unusable results, it is not the tool that is at blame, but the person who is using it incorrectly..
🤗 Transformers.js v4 preview is now available! npm i @huggingface/transformers@next And its huuuuge! I don't think a single line of code remained untouched😄. New build system, monorepo structure, and tons of new features!
We rearranged our apartment a bit and I had to move my office. A good time for some new paint and a few new plants🥰.
Over the Christmas holidays, I built three apps to replace my Google Sheets workflows: - 𝗙𝗶𝗻𝘁𝗿𝗮𝗰𝗸: A budget tracker that imports bank transactions and allows us to categorize them - 𝗣𝘂𝗺𝗽𝗲𝗹: A gym tracker - 𝗛𝗮𝗯𝗶𝘁𝘀: A daily habit tracker
Definitely not an easy post to write. A friend of mine died in an avalanche a few days ago. He was a professional snowboarder who knew the mountains intimately.
🚀 Transformers.js v4 update: We've migrated from webpack to esbuild. Results: - Build time: 2s → <200ms (10x faster) - Bundle size: 10-50% smaller Sometimes it's not about new features. It's about building a more robust architecture that makes everything else possible🔥. github.com/huggingface/...
[v4] Switch build system to use esbuild by nico-martin · Pull Request #1466 · huggingface/transformers.js
github.com
Passkeys aren't just more secure because of the tech, but also because of UX. Normally I would max out session tokens to avoid typing email/pw. Now with passkeys? Login is one click. Reauthenticate every time? Absolutely no problem..
I made myself a little Christmas present: a workout tracker that runs 100% offline in your browser. I's built entirely with Claude Code, zero manual coding, just conversations. Every component, every database query, even the deployment config. pumpel.nico.dev
Last week Google released FunctionGemma, a tiny open model for tool calling. And of course it does work in Transformers.js🔥! The remarkable part? This example uses the q4 variant, which only works because it's fine-tuned (locally) to the specific functions!
This is a todo app. BUT, it was entirely coded by Qwen3 Coder 30B running locally via LM Studio on my Mac M3. All within Claude Code. One prompt, done. No cloud. No API costs. Just a really good 30B model and agentic coding.
I'm pretty stoked to release a little side-project. The Language Model Polyfill, a cross platform polyfill for the Prompt API based on Transformers.js and the Qwen3-4B model: github.com/nico-martin/...
GitHub - nico-martin/language-model-polyfill
Contribute to nico-martin/language-model-polyfill development by creating an account on GitHub.
github.com
Can we take a moment to appreciate how good the Islands theme looks in @jetbrains.com 2025.3? The design team absolutely nailed it🥰
After an incredible year of conferences, I'm looking at 2026 with a fresh slate✨. Which conferences should be on my radar? Always happy to connect with organizers! I love creating and speaking about hands-on demos around WebML, Transformers.js, and bringing AI to browsers.
Did you know that you can use Transformers.js with the @vercel.com AI SDK? Huuuge shout out to www.jakobhoeg.dev for building an on-device AI provider! github.com/jakobhoeg/bu...
GitHub - jakobhoeg/built-in-ai: TypeScript library for using in-browser AI models with the Vercel AI SDK, with support for seamless fallback to server-side models
TypeScript library for using in-browser AI models with the Vercel AI SDK, with support for seamless fallback to server-side models - jakobhoeg/built-in-ai
github.com