Distributing AI with Python in the Browser: Edge Inference and Flexibility...
summary
Demos deliver the specifics: a Python runtime uses PyScript and Pyodide to expose Python tools in the page, a local Quen 3.8B model and a Quen 2.5 browser model run interchangeably, and Agent Studio exports an agent folder that the page mounts to edit files via mount local folder via Chrome. The prototype shows tool calls, a breakpointable agent loop, client‑side inference working offline, and tradeoffs like 700–900MB downloads and performance limits.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:38Defining agents
- 7:32Live demo setup
- 12:11Local model demo
- 15:39File handling demo
- 21:20Anaconda Agent Studio
- 23:49Browser agents demo
- 26:42Browser agents pros and cons
- 28:49Summary
related talks
31mBrowser-based AI workflows in Jupyter
53mAI Didn’t Kill the Web, It Moved in!
02mDiscover Web AI: Client side Agents, Gen AI, and machine learning in the browser
45mThe marvelous world of Python in Browser: a nice surprise
16mUnleash the superpowers of the AI PC
46mThe Ultimate Guide to Browser-Based AI: Privacy-First, Offline-Ready and Future-Proof
29mClient side Web AI Agents Building smarter user experiences for a future age
33mLocal Models: How to Run AI in the Browser