1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
const axios = require('axios');
const api_key = "YOUR API-KEY";
const url = "https://api.segmind.com/v1/gemini-omni-flash";
const data = {
"prompt": "Cinematic slow-motion shot of powerful ocean waves crashing against dramatic coastal cliffs at golden hour, sea spray catching warm sunlight, seabirds gliding in the distance, smooth aerial drone camera movement, photorealistic, highly detailed, 4k.",
"duration": 5,
"aspect_ratio": "16:9",
"seed": -1
};
(async function() {
try {
const response = await axios.post(url, data, { headers: { 'x-api-key': api_key } });
console.log(response.data);
} catch (error) {
console.error('Error:', error.response.data);
}
})();Attributes
Text describing scene, motion, style, and audio. Add camera and lighting cues for cinematic shots.
Clip length in seconds: 3, 5, or 10. Use 10 for fuller scenes.
Allowed values:
Output orientation, 16:9 or 9:16. Pick 9:16 for Shorts, 16:9 for widescreen.
Allowed values:
Integer from -1 to 999999999 for reproducibility. Use -1 random, fix to repeat.
min : -1,
max : 999999999
Optional reference video URL to restyle or edit. Beta; may be intermittent.
Optional reference image URLs (image-to-video / multi-image reference). Up to 14 images.
To keep track of your credit usage, you can inspect the response headers of each API call. The x-remaining-credits property will indicate the number of remaining credits in your account. Ensure you monitor this value to avoid any disruptions in your API usage.
Gemini Omni Flash — Multimodal Text-to-Video and Image-to-Video Model
What is Gemini Omni Flash?
Gemini Omni Flash is Google DeepMind's first model in the Gemini Omni family, a multimodal video model that creates and edits high-resolution video with synchronized native audio. Unlike single-task text-to-video generators, Omni Flash reasons natively across text, image, and video inputs, so the output is grounded in Gemini's knowledge of physics, geography, culture, and language. On Segmind you can call it as a fast, synchronous API: send a prompt (and optionally a reference image or video), and receive an MP4 directly in the response. It generates short clips of 3, 5, or 10 seconds in 16:9 or 9:16, making it ideal for social shorts, ads, explainers, and rapid creative iteration.
Key Features
- Text-to-video generation with synchronized audio from a single prompt.
- Image-to-video — animate a still or anchor a character from a reference image.
- Reference video (beta) — restyle or transform an existing clip.
- World knowledge for physically and contextually accurate scenes.
- Native audio sync so on-screen action and sound line up.
- Flexible output at 3/5/10 seconds, landscape or portrait, with a seed for reproducibility.
Best Use Cases
Gemini Omni Flash fits creators, marketers, and developers who need polished video fast. Use it for YouTube Shorts and TikTok hooks (9:16), product demos and e-commerce clips, explainer and education videos that benefit from accurate world knowledge, brand mood films, storyboards, and style tests. Its image-to-video mode turns product shots, illustrations, or portraits into moving scenes, while the prompt-first workflow makes A/B creative testing cheap and quick. Because output ships with synchronized audio, it removes a separate voiceover or sound-design step for many short-form jobs.
Prompt Tips and Output Quality
Write descriptive prompts that name the subject, action, camera movement, lighting, and style — for example, cinematic, slow-motion, golden hour, aerial drone shot. Specify how on-screen text should appear and sync with the action for explainers. For consistent characters, pass a reference image. Keep durations at 10 seconds for fuller scenes and 3 seconds for quick loops. Set a fixed seed to reproduce a look. Note current limits: maintaining complete consistency across heavy edits, complex motion, and perfectly accurate text can still be a challenge.
FAQs
Is Gemini Omni Flash a replacement for Veo? It replaced Veo in the Gemini consumer app but runs alongside Veo for production work; Veo still leads on the highest resolutions and longer extensions.
What inputs does it accept? Text prompts, an optional reference image, and an optional reference video (beta). Audio input is not yet supported, though audio is generated alongside the video.
How long can clips be? 3, 5, or 10 seconds. Google describes the 10-second cap as a deployment choice rather than a model limit; longer pieces are stitched from multiple generations.
What aspect ratios are supported? 16:9 for widescreen and 9:16 for mobile and Shorts.
Does it generate audio? Yes — every clip includes synchronized native audio.
Is the output watermarked? Yes — Google embeds SynthID watermarking and C2PA Content Credentials to mark AI-generated content.
Popular Models
