# Gemini Omni 1.1 Flash - Core.Today AI API

> Google's fast multimodal video generation and editing model with native audio, using the Interactions API. Text-to-video, image-to-video, keyframe interpolation, reference-to-video, and video editing at 360p-4K.

- **Provider**: Google
- **Model ID**: google/gemini-omni-1.1
- **Category**: Video Generation
- **Credits**: 3490 per 10s 720p video — billed per output second (360p 116/s, 720p 349/s, 1080p 535/s, 4K 1046/s); edit mode bills the source video's length
- **Speed**: Fast
- **Quality**: High

## Features
- Native synchronized audio on every video — dialogue, sound effects, ambient sound
- Video editing: describe the change in the prompt and the rest of the video is preserved
- Keyframe interpolation between a start and end frame — orbits, zooms, timelapses, seamless loops
- Reference images guide subject and style without being used as literal frames
- 360p fast-draft mode up to 4K final output, 16:9 or 9:16

## Use Cases
- Editing existing footage with a text instruction ('make the sky stormy')
- Videos with dialogue and sound effects prompted directly in the text
- Smooth transitions between two keyframe images
- Character-consistent videos from reference images
- Cheap 360p drafts before a 4K final render

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - Text prompt describing the video — scene, camera movement, lighting, mood, and audio. Put anything to avoid in the prompt too (e.g. 'no dialogue')

### Optional
- **image**: string - Starting image to animate into a video. When provided without last_frame, the video is generated from this image
- **last_frame**: string - Ending image. Together with image, the model interpolates a smooth transition between the two frames
- **reference_images**: array (default: []) - Reference images that guide the subject and style (not used as literal frames) — e.g. specific characters or objects
- **video**: string - An existing video to edit. Describe the change in the prompt; the model preserves the rest. Billing uses the source video's measured length
- **resolution**: string (default: 720p) - Output resolution. 360p is a fast draft mode. Billing per output second: 360p 116, 720p 349, 1080p 535, 4k 1046 credits
  Options: 360p, 720p, 1080p, 4k
- **aspect_ratio**: string (default: 16:9) - Video aspect ratio. Ignored when editing an existing video
  Options: 16:9, 9:16

## Examples

### Text-to-Video with Ambient Audio
Generate a cinematic clip with prompted soundscape
```json
{
  "model": "google/gemini-omni-1.1",
  "input": {
    "prompt": "A cinematic drone shot through misty pine mountains at sunrise, gentle wind and birdsong. No dialogue.",
    "resolution": "720p",
    "aspect_ratio": "16:9"
  }
}
```

### Video Editing
Edit existing footage while preserving everything else
```json
{
  "model": "google/gemini-omni-1.1",
  "input": {
    "prompt": "Make the sky stormy with distant thunder, keep everything else the same",
    "video": "https://example.com/source.mp4",
    "resolution": "720p"
  }
}
```

### Keyframe Interpolation
Smooth transition between a start and end frame
```json
{
  "model": "google/gemini-omni-1.1",
  "input": {
    "prompt": "Smooth continuous timelapse from sunrise to sunset over the lake",
    "image": "https://example.com/sunrise.jpg",
    "last_frame": "https://example.com/sunset.jpg"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Prompt the audio too — dialogue in quotes, sound effects and music described explicitly; use 'no dialogue' or 'no music' to suppress
- Draft at 360p (a third of the 720p rate), then re-render the final at 1080p or 4K
- Add 'in a single continuous shot' or 'no scene cuts' for one unbroken take
- Time events with phrases like 'after 3 seconds, a bird flies in' or [0-3s]/[3-6s] timecodes
- Edit mode is billed by the source video's length — trim the source before editing long footage

## Documentation
https://replicate.com/google/gemini-omni-1.1