Video to video models and APIs.
Work on footage you already have: edit or restyle a clip, extend it, generate from reference videos or add word-timed captions.
How to pick a video to video model
These models start from footage you already have. Some edit or restyle it, some use it as a reference for a new shot, and some handle finishing work like captions and file conversion.
Edit a clip with an instruction
Pruna P Video Edit and Gemini Omni Flash R2V change a short clip from one instruction. Seedance 2.0 Mini V2V edits from up to 3 reference videos and a prompt.
New shots from reference footage
Seedance 2.5 Reference-to-Video keeps the motion and timing of a reference clip while you change the subject or the style. Wan 3.0 R2V and MiniMax H3 R2V keep characters and style consistent from shot to shot.
Remove an object
VOID by Netflix removes an object from a video and fixes the shadows, reflections and motion it leaves behind.
Finishing
Video Caption burns word-timed subtitles into a video, MMAudio adds sound effects that match the picture, and Fast Video Converter re-encodes to MP4, MOV, WebM and other formats.
Frequently asked questions
How long can my input video be?
Most of these models work on short clips. The limits are listed with each model’s parameters. Split longer footage into scenes first.
Is the original audio kept?
It depends on the model. Pruna P Video Edit keeps the audio of your clip, and several reference models can add synced sound. Others return video only.
How do I send a video to the API?
Upload the file in a multipart request or pass a public URL. Wiro downloads URL inputs itself, so the link has to be reachable from the internet.




