Agnes Video Generator
Idea in, finished video out: Agnes Video Generator writes the script, generates every scene, voices the narration, burns in the subtitles, and stitches the final MP4 from a single prompt. Single-shot clip mode turns one prompt into one video, with text-to-video, image-to-video, and keyframes options plus control over duration, resolution, seed, and negative prompt. Multi-shot story mode runs the full ten-step pipeline from story and character reference to finished film, Article Video splits pasted text into timed scenes and narrates each one, Digital Human builds a presenter from a generated or uploaded image, and Poetry Video recites a classical poem line by line over generated scenes. Narration voices are grouped by the interface's 22 languages with adjustable speed, and subtitles follow word-level timing with your choice of font, color, size, position, stroke, and background. Upload character or scene references to keep faces consistent, set end frames per scene, and pick keyframes, transition-frame, or independent chaining. Every pipeline step checkpoints, so an interrupted task resumes from where it stopped, and the task list lets you create, view, resume, and stop jobs. Add several API keys to rotate through rate limits, or swap the script writer for any OpenAI-compatible or Anthropic-compatible model. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. MIT licensed.
Deploy