2026/05/14 We updated some items and descriptions to more intuitive expressions.
Use Ltx-2.3 to generate videos with audio from images and prompts. The models used are as follows:
・ltx-2.3-22b-distilled-1.1_transformer_only-kj(https://www.seaart.ai/ja/models/detail/d7ff78le878c73br91ig)
Options
・There are 2-pass and 1-pass modes. Officially, 2-pass is recommended. In 2-pass, the first pass generates a half-size video, and the second pass upscales it by 2x.
・You can use Qwen3-VL-4B to expand prompts.
