Gemini Omni 1.1 Flash Reference to Video: AI Image-to-Video Generator

How to Use Gemini Omni 1.1 Flash Reference to Video

1. Gather your reference media Up to 7 reference images (image_urls) and up to 3 reference video clips (reference_video_urls, each at most 3 seconds long).

2. Write a prompt that addresses each reference by tag Reference media is addressed in the prompt in list order: the first image is <IMAGE_REF_0>, the second <IMAGE_REF_1>, and so on; the first video is <VIDEO_REF_0>, etc. Example: 'A cat inspired by <IMAGE_REF_0> walks through the setting in <VIDEO_REF_0>.'

3. Choose duration, resolution, and aspect ratio Duration is 3–10 seconds (default 8). Resolution is billed per second: $0.048/s at 360p, $0.16/s at 720p, $0.24/s at 1080p, $0.48/s at 4K.

4. Submit and poll

   curl -X POST https://api.muapi.ai/api/v1/gemini-omni-flash-1-1-reference-to-video \
     -H "Authorization: Bearer YOUR_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{
       "prompt": "A cat inspired by <IMAGE_REF_0> walks through the setting in <VIDEO_REF_0>.",
       "image_urls": ["https://example.com/cat.jpg"],
       "reference_video_urls": ["https://example.com/setting.mp4"],
       "resolution": "1080p"
     }'
   

   Then poll GET /api/v1/predictions/{request_id}/result until status is completed.

5. Prompt tips

   • Keep the tag order matched to the URL list order — <IMAGE_REF_0> is always the first image_urls entry.
   • Describe how each reference should be used (subject, setting, motion) rather than just naming the tag.
   • Reference videos longer than 3 seconds should be trimmed before submitting.