1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
const axios = require('axios');
const fs = require('fs');
const path = require('path');
async function toB64(imgPath) {
const data = fs.readFileSync(path.resolve(imgPath));
return Buffer.from(data).toString('base64');
}
const api_key = "YOUR API-KEY";
const url = "https://api.segmind.com/v1/wan2.6-i2v-flash";
const data = {
"prompt": "The wind blows through her hair and light dress as ocean waves crash against the rocks below. She slowly turns her head to look toward the camera, strands of hair drifting in the breeze. Warm golden-hour sunlight, gentle cinematic slow push-in.",
"image": "toB64('https://segmind-resources.s3.amazonaws.com/input/wan2.6-i2v-flash-input.jpg')",
"audio": true,
"audio_url": "toB64('null')",
"resolution": "720P",
"duration": 5,
"prompt_extend": true,
"shot_type": "single",
"watermark": false
};
(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
Describe subject motion, camera moves and mood. Be specific; short clear prompts work best.
First frame the video starts from; JPG, PNG or WEBP URL. Use sharp images.
Toggle synchronised soundtrack; affects pricing. Enable for dialogue, disable for silent clips.
Optional driving audio URL, 3-30s WAV or MP3. Supply to lip-sync motion to speech.
Output resolution, 720P or 1080P; affects pricing. Use 720P for drafts, 1080P for final.
Allowed values:
Clip length in seconds, 2 to 15; affects pricing. Use 5 for social clips.
Content to exclude, up to 500 characters. List artifacts like blur or distortion.
Auto-rewrites and enriches short prompts. Keep on for brief prompts, off for exact wording.
Single continuous shot or multi-shot edit; needs Prompt Extend. Use multi for narrative sequences.
Allowed values:
Seed from 0 to 2147483647 for reproducibility. Fix a value while iterating prompts.
Adds an AI Generated watermark; default off. Enable for disclosure, disable for clean output.
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.
Wan 2.6 Image to Video Flash — Image-to-Video Model
What is Wan 2.6 Image to Video Flash?
Wan 2.6 Image to Video Flash is the speed-optimized variant of Alibaba Tongyi Lab's Wan 2.6 image-to-video model. Give it a single starting image and a motion prompt, and it animates that frame into a fluid clip of up to 15 seconds at 720p or 1080p. The Flash version is distilled from the full model to return results faster while keeping the smooth motion and visual consistency that define the Wan 2.6 family, making it a strong fit for rapid iteration, batch generation, and prompt testing.
What sets it apart is native audio-visual synchronization. The model can generate a matching soundtrack — dialogue, sound effects, and ambient atmosphere — in the same pass as the video, so there is no separate dubbing step. You can also supply your own audio track and have motion and lip movement sync to it.
Key Features
- Image-to-video animation: Turns one still image into a coherent 2 to 15 second clip, using the input as the visual anchor for lighting, composition, and subject.
- Native audio in one pass: Generates synchronized sound effects, ambience, and speech with lip-sync; toggle audio off for silent clips.
- Custom driving audio: Upload a 3 to 30 second WAV or MP3 so motion and lips follow your own voiceover or music.
- 720p and 1080p output: Choose faster 720p drafts or 1080p for final delivery, both at smooth 24 fps.
- Multi-shot storytelling: Enable multi-shot mode to auto-split a prompt into several coherent scenes with consistent details.
- Prompt extend and negative prompt: Built-in prompt enhancement enriches short prompts, while a negative prompt keeps unwanted content out.
- Seed control: Fix a seed for more repeatable iteration.
Best Use Cases
Wan 2.6 Image to Video Flash shines wherever speed and volume matter. Marketing and social teams can animate product shots, flat-lays, and campaign stills into scroll-stopping content for TikTok, Instagram Reels, and YouTube Shorts. Because it returns quickly, it is well suited to A/B testing ad hooks, generating multiple variants of one prompt, and powering in-app video features. Artists and illustrators can bring a single key frame or piece of concept art to life, while educators can turn diagrams and reference images into short explainer clips. The multi-shot mode extends it to compact narrative sequences and product showcases where character and scene details need to stay consistent across cuts.
Prompt Tips and Output Quality
Start with a clean, high-resolution, well-lit image — the model amplifies whatever you feed it, so sharp subjects give better results than dark or compressed inputs. Treat the image as the anchor and the prompt as the motion and mood: name the subject, the action, the camera move, and the lighting. Something specific like "the person turns their head left while smiling, then looks back at the camera, slow push-in" outperforms "make it move." Keep prompts short and clear for image-to-video, fix a seed while you iterate so changes come from your edits rather than randomness, and use the negative prompt to suppress blur or distortion.
FAQs
How long can the videos be? Clips run from 2 to 15 seconds in a single generation, long enough for social posts, product demos, and short narrative arcs.
What is the difference between Flash and standard Wan 2.6? Flash is distilled for lower latency, trading some of the slower, more detailed processing for faster turnaround while keeping the core capabilities.
Does it generate audio automatically? Yes. It produces synchronized audio in the same pass, including sound effects and ambience, and you can also upload your own driving audio.
What resolutions and frame rate does it support? It outputs 720p or 1080p at 24 fps; the model preserves your input image's aspect ratio and scales pixels to the chosen resolution.
Can I control the exact output? A prompt, seed, and reference image guide the result, but the same seed does not guarantee identical frames, and fine details like faces, hands, text, and logos may shift during motion.
How do I create a multi-shot video? Enable multi-shot mode with Prompt Extend on, or describe the shot structure directly in your prompt using scene or timestamp cues.
Popular Models
