Create 2–30 second AI videos from text, first and last frames, or multimodal references in 480P, 720P, or 1080P with audio and no watermark.