From a prompt
Describe the shot: who is in it, what moves, where the camera goes.
Describe a shot, or start from one of your pictures. Nook renders the clip on your card with LTX-Video and Wan 2.2: a presenter talking to camera, a product turning, a slow push-in. No queue, no credits, no watermark.
Windows 11 · NVIDIA 8 GB or more recommended · No account, no cloud, no upload
Start from a sentence, or from a picture you made or took. The clip renders on your GPU and is saved in a folder on your PC.
Describe the shot: who is in it, what moves, where the camera goes.
Give it one of your pictures and say how it should move. Good for product shots and for bringing a render to life.
No credits and no queue, so render the shot again until it is right.
Nook sizes each model to your card before it downloads.
Fast, and the model behind the clip on this page: ten seconds at 1080p.
Text to video and image to video from the open Wan family, strong on motion and camera moves.
| Graphics card | Clips |
|---|---|
| 8 GB | Short 480p clips. |
| 16 GB and up | 720p and 1080p, and longer shots. |
Yes. Nook runs both on your own GPU on Windows 11: install, pick a model, describe the shot. Short 480p clips fit an 8 GB card; 1080p and longer shots want 16 GB and up.
No. The clip renders on your card, so there is no queue, no credit counter and no watermark.
Yes. Nothing is uploaded: the model runs on your GPU and the clip is saved in a folder on your PC.
Yes. Start from any still, including one you made with Nook's image generator, and describe how it should move.
Yes, short clips. 480p clips fit an 8 GB card; 720p, 1080p and longer shots want 16 GB and up.
Only to download a model. After that clips render offline.
One app for the AI you run at home: a coding agent, pictures, video, and speech to text with Whisper. All of it on your GPU.
A Windows desktop app: one download, no account, no admin rights, and it keeps itself up to date.