Emotion Analysis API

The emotion analysis endpoint processes images and video to produce a seven-emotion timeline — enjoyment, sadness, anger, fear, surprise, disgust, and contempt — grounded in the Facial Action Coding System (FACS). Analysis is asynchronous: submit a job, then poll or webhook to retrieve results.

Job Lifecycle

Submit — POST an image URL, base64 image, or video URL to create an analysis job. Returns a job ID immediately.
Queued — Job is accepted and waiting for a processing slot.
Processing — Frames are being analyzed.
Completed — Results are available; retrieve the timeline via the job status endpoint.
Failed — Processing failed; not billed. Retry with the same or a corrected payload.

Results Format

For images: a single per-face emotion score vector (0–1 for each of the seven emotions). For video: a per-frame or per-second timeline of emotion scores, plus aggregate summaries (modal intensity, overall distribution). Results can be downloaded as JSON or CSV.

Pricing

Image analysis: $0.002 per image. Video analysis: $0.08 per minute (rounded up to next full minute). Failed jobs are not charged.

Face Detection | API Pricing | Tutorial: Emotion in Video | Developer API overview