Detect faces/speakers and write layout + keyframes
Synchronous (10 s – 2 min). Two stable faces (> 60 % of the time) → one split keyframe; otherwise vertical keyframes following the active speaker (ease: smooth); horizontal → full frame. Overwrites layout and keyframes of the edit for this orientation. Results are cached per clip bounds. 500 if the face model is missing on the server.
Request
- Base URL: https://cliponit.app
- URL: https://cliponit.app/api/clips/{id}/auto-reframe
- Auth: one of:
- HTTP bearer
- API key in cookie session
- OAuth 2 (authorization code, authorization URL https://cliponit.app/oauth/authorize, token URL https://cliponit.app/oauth/token, scopes: clips)
Path parameters
Clip id (UUID)
Query parameters
vertical = 9:16 1080x1920 (TikTok/Reels/Shorts), horizontal = 16:9 1920x1080 (YouTube), square = 1:1 1080x1080 (feed), portrait45 = 4:5 1080x1350 (Instagram feed). auto-reframe supports vertical|horizontal only.
Which edit/export: vertical (9:16, default), horizontal (16:9), square (1:1) or portrait45 (4:5). Each clip has one independent edit per orientation.
Response
Reframe result
Example response
{
"keyframes": [
{
"t": 0,
"layout": "vertical",
"zones": [
{
"x": 0.3,
"y": 0,
"w": 0.4,
"h": 1
}
],
"ease": "cut"
}
]
}Changes
No recorded changes to this operation across all 1 revision of this API.