Video API integrations

Same endpoint for 19 models. Each guide includes tested request shapes and the polling loop. There is no video-completion webhook yet.

How every integration works

Whatever the language or tool, a VideoGenAPI integration is two HTTP calls. Submit a job with POST /api/v1/generate, then poll GET /api/v1/status/{id} until the status is completed and download video_url. Authentication is a single Authorization: Bearer header, and switching between Kling 3, Veo 3, Seedance or MiniMax H3 only changes the model field.

POST https://videogenapi.com/api/v1/generate
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json

{"model": "kling-3", "prompt": "A dancer spins in a sunlit studio", "duration": 5}

GET https://videogenapi.com/api/v1/status/{generation_id}

Which integration to pick

  • Python: scripts, data pipelines, batch jobs and Django or FastAPI backends. Start with the Python guide or the longer Python tutorial.
  • JavaScript / Node: Next.js API routes, serverless functions and workers. See the JavaScript guide; keep the key server-side.
  • n8n, Make, Zapier: no-code automations such as product videos from a spreadsheet or daily social clips. The n8n guide includes an importable workflow and the polling loop.
  • Claude, ChatGPT, Cursor: let an AI agent generate video from a conversation through the MCP server, with OAuth instead of pasting your key.

Production checklist

  • Store the API key in an environment variable or secrets manager (authentication).
  • Persist each generation_id before polling, and poll every 10 to 15 seconds with a 10-minute timeout (polling patterns).
  • Back off on 429 responses using details.seconds_until_reset (error codes, rate limits).
  • Download finished videos to your own storage if you need them long-term.
Reviewed by Enzo Lanne · Founder, VideoGenAPI ·