Reframe a video to a new resolution
Reframe a video to a new resolution. The model fills in the missing areas with generated content that matches the original. On completion, the result contains video_url, the source video reframed to the requested resolution.
Responds immediately with the job id and created_at timestamp. Poll GET /v2/video-to-video-reframe/{id} until the status is completed.
Billed per second of input video, by output resolution. See Pricing.
Authentication
Request
Input video in HTTPS URL or base64 data URI format. See Input Formats for supported formats and codecs.
Max 60 seconds and 1800 frames (e.g. 60s at 30fps, or 30s at 60fps).
Target output resolution. The source video is extended to fill this frame. Pass one of the values below; each maps to an aspect ratio:
Where the source video lands on the output frame: an Anchor such as { "anchor": "top" } to fit and align it, or a Rectangle such as { "x": 0, "y": 0.35, "width": 1 } to set its position and size exactly (see both below). The whole source is always kept and never stretched; the model generates the rest of the frame. Omit to center it at the largest size that fits.