One Click Videos · Developers

One Click Videos with Gemini

Connect Gemini function calling to a video workflow executed by your application.

Define three application functions

Function in your app JEMSU request
get_video_credits GET /credits
create_video POST /videos
get_video_status GET /videos/{jobId}

These are suggested function names for your own adapter, not hosted JEMSU MCP tools.

Keep the executor server-side

  1. Define the function input schemas from the OpenAPI contract.
  2. Let Gemini propose a structured brief from the user’s request.
  3. Validate the arguments and the user’s authorization before your server calls JEMSU.
  4. Return the job ID as the tool result.
  5. Poll in your worker and provide the completion result when available.

A useful tool result

{
  "jobId": "RETURNED_JOB_ID",
  "status": "submitted",
  "nextAction": "Check get_video_status using this jobId."
}

Never include the API key in the tool result or prompt. Record the job so an interrupted assistant exchange does not create a duplicate.

Function calling lets a Gemini application connect model-selected actions to external code. Use Google’s official function-calling documentation for the current SDK-specific setup.