Your video. A new cast, setting, or product. Add reference images and describe what to change while using the original motion as a guide.
An existing 15-second Wan 3.0 Prime result from our Bulin 47 template, with its actual input photo. Custom edits will vary.
Video Recast is a video-to-video AI tool. Start with a clip, add images of the people, places, clothing, or products you want to use, and describe the changes. The source video guides the movement and camera; the images guide the new appearance.
Unlike a fixed template, you choose the reference clip and the edit. This creates a new video rather than modifying individual pixels, so check the result for changes to faces, objects, and background details.
Choose a short clip with the action and camera movement you want. A clear subject and a simple shot make it easier to describe the edit.
Upload the new characters, setting, product, or outfit. Turn on Prompt if you want to assign images or describe specific changes; otherwise, use the default instructions.
Choose your settings and review the credit cost before generating. Check the finished video, then download it from your result or history.
Use these instructions as starting points. Keep the target clear, and review the generated result before sharing it.
Use your images to guide faces, hair, clothing, and body shape while the source guides the performance. Assign each person explicitly in a multi-person clip.
Replace the person on the left in Video 1 with the person in Image 1. Keep the other person, setting, and camera movement.
Try a different location around the original action. Describe the scene to replace and use a reference image to show the new surroundings.
Move the performance in Video 1 to the location in Image 1. Keep the same performers, actions, framing, and timing.
Point to a specific object and provide a clear product image. Inspect labels, proportions, and contact with hands in the result.
Replace the bottle held in Video 1 with the bottle in Image 1. Keep the person's appearance, hand movement, and background.
Guide a wardrobe change without asking for a new character. Include the garment's shape, material, and fit in the reference or prompt.
Dress the main performer in Video 1 in the jacket from Image 1. Keep their face, hairstyle, body proportions, and movement.
No. A face swap focuses on a face. Video Recast uses a source video, reference images, and instructions to guide broader changes, including a person's appearance, clothing, setting, or an object. It generates a new video, so results can differ from the source.
There is no pixel-perfect preservation guarantee. State both the intended change and the elements to keep. The model can still alter details, especially around occlusion, fast movement, faces, hands, text, and products. Review the full video before using it.
Upload a reference video, then clear images of the characters or elements you want to introduce. Follow the file, duration, and image limits shown in the tool. Use assets you have permission to upload and edit; do not include private information you do not want processed.
Prompt is optional and off by default. Turn it on to name the target in Video 1, assign images by number, and describe what to keep. For example: replace the left performer with the person in Image 1; preserve the other performer and the camera movement. With Prompt off, the tool uses default recasting instructions instead of your saved custom text.
Trending templates already include a selected clip and role instructions. Video Recast lets you supply your own video, reference images, and prompt. Choose a template for a ready-made format, or use this tool for your own scene.
The tool shows the required credits before you submit, based on the supported settings and source video. Review that amount before generating. A new generation is a separate attempt; a different result is not guaranteed.
Our Trending templates pair a selected reference clip with role instructions, so you can start with your photos.
