Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
Our customers are pushing the frontiers of video generation. 🍋📽️ Real-time video calls with AI characters, powered by a custom video diffusion transformer. Instead of requiring a custom model trained per character, @LemonSliceAI (YC W24)'s zero shot model only needs a single https://t.co/Z2qyatAe30
🍋@LemonSliceAI just launched Character AI with video. It’s an audio-video AI model that can instantly create any character from just a single photo, in any style. Every pixel on the screen is generated at 25fps by a transformer model in real-time. https://t.co/XOBeolGdTv
First time AI has made me laugh out loud. Congrats @LemonSliceAI on the launch! Try uploading an image yourself and watch it come to life. No waitlist, works on everything from cartoons to photorealistic images, and every pixel on the screen is generated at 25fps by a transformer https://t.co/rCb54RMwpj
Lemon Slice Live, developed by LemonSliceAI and launched in April 2025, is a real-time audio-video AI model that enables users to video chat with any character by uploading a single photo. The model uses a transformer architecture to generate every pixel on the screen at 25 frames per second, supporting a wide range of image styles from cartoons to photorealistic images. This zero-shot model does not require custom training per character, allowing instant creation and interaction with AI-generated characters. The technology has been highlighted for its innovative approach to video generation and real-time communication.
