Lip Sync

Lip Sync picks a deliberately weird stage for an audio experiment: two animated mouths discussing an artist-curated question. Choose a subject around art, craft, food, or travel, then listen as generated voices turn it into a compact back-and-forth.

What is Lip Sync?

The premise is closer to a conversation starter than a reference desk. More than 100 prepared questions supply the raw material, while Gemini produces a two-speaker script and a Google audio model gives it voices. The disembodied-mouth presentation keeps the machinery visible and a little ridiculous, which is part of the charm. What comes back is generated on demand, so the pleasure lies in hearing an unexpected angle rather than treating every sentence as settled fact.

What you can do there

• Choose a curated cultural question
• Listen to two generated voices discuss it

Why we picked it

Plenty of generative tools hide behind a blank prompt box. Lip Sync gives the model a tightly framed role, a curated question set, and a visual gag you understand in one glance. That constraint makes the experiment easier to approach and more distinctive than an open-ended chatbot.

How to get the most from it

Pick a topic you already know enough about to spot a shaky claim, then treat the exchange as fuel for a real conversation. Try neighboring questions to see whether the two voices find genuinely different angles or merely swap agreeable sentences.

Good to know

Sound is the main event, and the page uses animated motion. Generated discussions can be wrong or oversimplified, so verify any factual point elsewhere. Caption support, device behavior, account requirements, and accessibility controls are not documented on the project page.

Who made it, and when?

Gaël Hugo is the credited creator or organization. The earliest supported launch date we found is November 1, 2024.