Inference
Describe
Generates descriptions for video content based on a given prompt.
POST
/
describe
/
This feature is in beta and may change without notice.
Request
file
file
requiredThe video file to be described. Supported formats: .mp4, .avi, .mov, .mkv, .flv, .wmv
prompt
string
requiredThe description or prompt of what to look for in the video.
Response
The API returns a JSON object containing the structured description of the video based on the given prompt.
Notes
- The video file size is limited to 5MB per request. For larger files, consider using the File Tools API to pre-process the input.
- The API uses the
video-descriptor-v1
model for video description.
Was this page helpful?