Audio-Visual Vibe Coding with Qwen 3.5 Omni: Spoken Specs to Web App

Melvin Vivas · X video post · 2026-03-30 · 2:47 · 108 views · Open on X

Topics: AI Dev Tools & Productivity, LLM Fundamentals, Industry Trends & Job Market · Level: beginner

Summary

This is a reshared demo from Alibaba's Qwen team. A user talks to the Qwen 3.5 Omni model and points at parts of a layout to describe a product and scenery showcase webpage, and the model generates the HTML. The user then asks out loud for refinements, such as detail pages, a buy button and a like button. The demo shows how a multimodal model can turn spoken, visually grounded instructions into working front-end code.

Key points

• Qwen 3.5 Omni takes audio and visual input together. The user speaks while pointing at areas of a layout ('in this section... display an airplane'), and the model maps each instruction to the right region.
• First prompt: build an HTML page with two toggle buttons, Product and Scenery. Each one swaps the image grid between product photos and high-resolution landscape photos.
• The spec asks for unlimited images, real image URLs and a 'professional, high-end' layout, which shows how much one spoken prompt can cover.
• Second round: clicking a product image opens a detail page with a large hero image, the name, the price and a Buy button. The Buy button triggers a 'purchase successful' notification. The page also lists specs such as size, dimensions and origin.
• Clicking a landscape image opens a different detail page with the place name, city, elevation, a red heart 'like' button with a visual effect and a like-count notification, plus a description.
• Pattern: build in rounds. Start with a basic layout, then add interactions and different detail pages for each content type in follow-up spoken prompts.

Resources mentioned

• Qwen 3.5 Omni · tool · qwen.ai · free
  Alibaba's omni-modal Qwen model that accepts audio and video input and can generate code from spoken, visually grounded instructions.
• Qwen (@Alibaba_Qwen) on X · person · x.com · free
  Official X account of Alibaba's Qwen team, which posts model releases and demos.
  Also in: Qwen-Audio-3.1: Alibaba's Five-Model Audio Stack (Melvin Vivas on X · notes), Demo: Qwen3.8-27B Running Locally with Pi and llama.cpp (Melvin Vivas on X · notes), Running AI Locally: Rebuilding AIBackends as a Python Library with Open Models (Melvin Vivas on X · notes)

Try this

• Build a product/scenery showcase webpage by voice: two toggle buttons that switch image grids, with clickable detail pages (product: price, Buy button, success popup, specs; landscape: city, elevation, like button with counter).