gemini-omni

Create Gemini Omni voice resources, character resources, and Flash Preview or multimodal text-to-video tasks through RunAPI. Use when the user asks an agent to create or manage Gemini Omni audio voices, character resources, or video. Default to the RunAPI CLI for one-off calls; use SDKs only when integrating RunAPI into an app or backend.

Gemini Omni On RunAPI

Create Gemini Omni voice resources, character resources, and text-to-video tasks through RunAPI. The default path for one-off agent tasks is the runapi CLI; SDKs are for application integration.

Critical: Integration Runtime

• Integration work (app, backend, worker, library, Rails service, Node service, Go service, webhook pipeline, or production codebase) uses the SDK integration path for the target language.
• One-off generation, editing, transformation, manual smoke tests, debugging, or user-requested CLI runs use the CLI path with the runapi binary. For full CLI-specific agent guidance, see https://github.com/runapi-ai/cli-skill.
• Never shell out to the runapi CLI as the production runtime integration layer.

SDK integration path

When integrating Gemini Omni into an app, backend, worker, library, Rails service, Node service, Go service, webhook pipeline, or production workflow, start by checking the current SDK package and official usage. Confirm install commands, client methods (create, get, run), request fields, response shape, and error classes before using CLI help or raw HTTP examples. Use a RunAPI SDK package:

• JavaScript / TypeScript: @runapi.ai/gemini-omni
• Python: runapi-gemini-omni
• Ruby: runapi-gemini-omni
• Go: github.com/runapi-ai/gemini-omni-sdk/go
• Java: ai.runapi:runapi-gemini-omni
• PHP: runapi-ai/gemini-omni

Variants

• Audio: create reusable voice resources with runapi gemini-omni create-audio.
• Character: create reusable character resources with runapi gemini-omni create-character.
• Video: use gemini-omni-flash-preview for prompt-only 720p generation, or gemini-omni-text-to-video for the existing multimodal workflow.

CLI path

The runapi binary is the one-off and manual testing runtime dependency. For full CLI-specific agent guidance, see https://github.com/runapi-ai/cli-skill. Run runapi auth status first. For agents and headless runs, prefer RUNAPI_API_KEY or import it into saved config.

Inspect the available commands and request fields with CLI help:

shell

runapi gemini-omni --help
runapi gemini-omni create-audio --help
runapi gemini-omni create-character --help
runapi gemini-omni text-to-video --help

Run a one-off request:

shell

runapi gemini-omni create-audio --input-file request.json

For video tasks, submit asynchronously and wait for completion:

json

{
  "model": "gemini-omni-flash-preview",
  "prompt": "A paper airplane glides through a sunlit studio.",
  "aspect_ratio": "16:9",
  "output_resolution": "720p"
}

shell

runapi gemini-omni text-to-video --async --input-file request.json
runapi wait <task-id> --service gemini-omni --action text-to-video

Available commands: create-audio, create-character, text-to-video.

Generated file storage

RunAPI-generated file URLs are temporary. Download and store generated images, videos, audio, or other files in your own durable storage within 7 days; do not treat returned URLs as long-term assets.

References

• Flash Preview model, pricing, and rate limits: https://runapi.ai/models/gemini-omni/flash-preview.md
• Provider comparison: https://runapi.ai/providers/google.md
• Full model catalog: https://runapi.ai/models.md