Transformers.js: Building Next-Generation WebAI Applications
summary
Transformers.js runs models 100% locally in the browser using WebGPU and WebNN and claims 1.7M unique monthly users with ~11M CDN requests. Models are quantized up to 8x; a 1.7B model hits ~160 tokens/sec on an M4 Mac. Demos show real‑time Whisper speech recognition, an 82M‑parameter Coco TTS, background removal, live video captioning, semantic embedding search and a drawing game — all on‑device with low latency.
AI-written summary · about
chapters (10)
- 0:00Transformers.js intro
- 2:06What is Transformers.js
- 4:34Optimizing models
- 8:28Usage growth
- 10:13Three-line usage
- 13:19Building web AI apps
- 15:03Key factors
- 16:08Real-time AI demos
- 22:07Multi-model voice assistant
- 23:57Future plans
related talks
26mDeploying PyTorch Models To the Browser and Beyond With Transformers.js
33mUsing Transformers.js to Bridge the Gap Between AI Research and
22mBoring but Effective: WebAI That Ships
57mMachine Learning with JavaScript
41mEscaping the Cloud: High-Performance AI in your Browser
33mLocal Models: How to Run AI in the Browser
39mMachine Learning Magic for Your JavaScript Application
29mSuperpowers for next gen web apps: Machine learning