Skip to Content
DocumentationVideo RenderingOverview

What is Video Rendering?

The Video Rendering API generates lip-synced talking-head videos. You provide:

  • A text script (or audio from a voice agent session)
  • An avatar image
  • Optional reference audio for voice cloning

Oshara processes the job asynchronously and returns a URL to the rendered video once complete. Jobs can also be published directly to YouTube.

Job lifecycle

POST /api/video-jobs/
        │
        ▼
  status: PENDING
        │
        ▼  (worker picks up job)
  status: PROCESSING
        │
     ┌──┴──┐
     ▼     ▼
COMPLETED  FAILED
     │
     ▼
processed_video_url (S3 signed URL)

Typical integration

1. Create a job → receive job ID
2. Poll GET /api/video-jobs/{id}/ every 5 seconds
3. When status == "COMPLETED", download from processed_video_url
4. (Optional) POST /api/youtube/upload/ to publish to YouTube
EndpointDescription
Video JobsCreate and poll rendering jobs
HLS StreamingGenerate and stream video in real time via HLS
YouTube IntegrationConnect a YouTube channel and publish finished videos

Rate limits and quotas

Video rendering is compute-intensive. Concurrent job limits depend on your plan. Contact hello@oshara.ai for enterprise quotas.

Last updated on
QueryVideo Jobs