Show HN: Infinity – Realistic AI characters that can speak

Infinity AI introduced a new tool called Infinity which generates realistic AI characters that can speak synchronously with audio input. The proprietary video diffusion transformer model was trained for approximately 11 GPU years or around $500k. Despite some limitations, including slow processing times, the model can handle multiple languages, learned some physics for realistic movement, and can animate different types of images. It can even handle singing, offering possibilities for creators to generate unique videos by simply typing a script.

• Infinity AI releases AI character tool using audio-driven video diffusion transformer.
• Infinity tool allows creators to script character dialogues for video generation.
• The model handles multiple languages and understands physical movements like dangling earrings.
• Capable of animating diverse visuals including paintings, sculptures, and singing characters.
• Known limitations include slow processing and difficulty with non-humanoid characters.