Gemini Omni Python SDK for RunAPI

The Gemini Omni Python SDK is the language-specific package for Gemini Omni on RunAPI. Use this package for voice resources, character resources, and multimodal video generation workflows when your application needs request bodies, task status lookup, and consistent RunAPI errors in Python.

Install

pip install runapi-gemini-omni

Quick start

from runapi.gemini_omni import GeminiOmniClient

client = GeminiOmniClient()  # reads RUNAPI_API_KEY, or pass api_key="sk-..."

# Reusable voice (returns immediately)
voice = client.create_audio.run(audio_id="kore", name="Narrator")

# Reusable character (returns immediately)
character = client.create_character.run(
    descriptions="A friendly robot guide",
    reference_image_url="https://cdn.runapi.ai/public/samples/portrait.jpg",
    body_reference_image_url="https://cdn.runapi.ai/public/samples/image.jpg",
)

# Text-to-video (create + poll until complete)
result = client.text_to_video.run(
    model="gemini-omni-flash-1-1",
    prompt="A paper airplane travels from dawn into dusk.",
    duration_seconds=6,
    first_frame_image_url="https://cdn.runapi.ai/public/samples/first-frame.jpg",
    last_frame_image_url="https://cdn.runapi.ai/public/samples/last-frame.jpg",
    aspect_ratio="16:9",
    output_resolution="360p",
)
print(result.videos[0].url)

character.images returns the portrait first and the optional full-body image second. A character created with both references consumes two of the seven reference units in a multimodal video request.

For gemini-omni-flash-1-1, first_frame_image_url cannot be combined with reference images, audio IDs, video clips, or character IDs. last_frame_image_url requires first_frame_image_url.

create_audio returns its result directly. create_character.run returns the terminal character result and follows an accepted Task when needed. text_to_video is asynchronous: use create to submit and return quickly, get to fetch the latest task state, and run to create and poll until completion.

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.

Language notes

Pass parameters as keyword arguments and catch the runapi.gemini_omni error classes. The available resources are create_audio, create_character, and text_to_video. Keep RUNAPI_API_KEY in the environment or your secret manager; never commit API keys or callback secrets.

Links

• Model page: https://runapi.ai/models/gemini-omni
• SDK docs: https://runapi.ai/docs/resources/sdks
• Product docs: https://runapi.ai/docs/api/gemini-omni/text-to-video
• Flash 1.1 pricing and rate limits: https://runapi.ai/models/gemini-omni/flash-1-1
• Flash Preview pricing and rate limits: https://runapi.ai/models/gemini-omni/flash-preview
• Full catalog: https://runapi.ai/models

License

Licensed under the Apache License, Version 2.0.