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
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/scail";
const data = {
"prompt": "a cat playing",
"reference_image": "toB64('https://segmind-resources.s3.amazonaws.com/input/a27216d9-5e6a-497e-a995-392f13e171e3-scail-ref-image.jpg')",
"input_video": "https://segmind-resources.s3.amazonaws.com/input/4c379a6a-f6af-445b-aa82-8baf4d3c67d4-scail-video.mp4",
"seed": 987654321,
"resolution": "576p",
"base64": 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 the scene or action to render. Example: 'a cat playing'.
Reference image URL to guide output. Use images matching prompt theme.
URL of video to process or enhance. Ensure it matches the scene style.
Set random seed for consistent results. Try 987654321 for varied images.
Choose video resolution. Use 720p for high-quality outputs.
Allowed values:
Output as base64 for integration. Set true for web embedding.
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.
SCAIL: Studio-Grade Character Animation Model
What is SCAIL?
SCAIL (Studio Character Animation with Intelligent Learning) is a cutting-edge AI model that generates professional-quality character animations from reference images and motion videos. Built on a diffusion-transformer architecture, SCAIL understands 3D-consistent pose representations to maintain structural accuracy and temporal coherence across frames. Unlike conventional animation models, SCAIL employs a novel full-context pose injection mechanism that captures spatial-temporal dynamics, enabling it to handle complex cross-identity motions while preserving anatomical correctness. Whether you're animating human performers or stylized anime characters, SCAIL delivers film and game production-ready outputs.
Key Features
- 3D-Consistent Pose Learning: Maintains anatomical accuracy and structural integrity across all frames
- Cross-Identity Motion Transfer: Apply motion from one character type to completely different character designs
- Temporal Coherence: Ensures smooth, natural animation flow without jarring transitions or pose drift
- Multiple Resolution Support: Generate animations at 480p, 576p, and 720p for various production needs
- Reference-Guided Generation: Uses static reference images to maintain character appearance consistency
- Motion Video Input: Accepts video input to guide character movement and action sequences
- Deterministic Output Control: Seed parameter enables reproducible results for iterative refinement
Best Use Cases
Film Production: Generate pre-visualization animations or background character movements for VFX-heavy scenes requiring consistent character performance.
Game Development: Create diverse NPC animations, cutscene sequences, and character idles while maintaining visual consistency across different character models.
Animation Studios: Rapidly prototype character movements, explore animation variations, or extend motion capture data to stylized characters.
Content Creation: Produce social media content, YouTube animations, or marketing materials featuring animated characters without extensive manual keyframing.
Virtual Production: Drive digital avatars or virtual presenters with reference-based motion while preserving brand-specific character designs.
Prompt Tips and Output Quality
Be Specific About Actions: Instead of "person moving," write "woman walking confidently forward while waving right hand." Detail improves pose accuracy.
Match Reference to Prompt: Ensure your reference image style (realistic vs. anime) aligns with your desired output aesthetic for best results.
Video Quality Matters: Clear, well-lit input videos with visible body movements yield more accurate pose extraction and transfer.
Resolution Selection: Use 576p for balanced quality and processing speed; upgrade to 720p for final production renders requiring fine detail.
Seed Experimentation: Start with default seed (987654321), then iterate with different values if subtle pose variations are needed.
Scene Context: Include environmental context in prompts ("dancing on stage," "running through forest") to help the model understand spatial relationships.
FAQs
Q: What file formats does SCAIL accept for reference images and input videos?
A: SCAIL accepts standard image formats (JPEG, PNG) for reference images and common video formats (MP4) via URL input.
Q: Can I use SCAIL for non-human characters like animals or creatures?
A: Yes, SCAIL handles diverse character types. The model's 3D pose learning adapts to various anatomical structures, though results are optimized for humanoid forms.
Q: How does SCAIL differ from traditional motion capture or keyframe animation?
A: SCAIL bridges the gap between the two. It transfers motion semantics from video to new character designs without requiring expensive mocap suits or manual frame-by-frame animation.
Q: What's the recommended workflow for iterating on animation results?
A: Start with clear reference images and motion videos, use consistent seeds for A/B testing, then refine prompts based on output. Adjust resolution last when satisfied with motion quality.
Q: Does the model maintain character identity when transferring complex poses?
A: Yes, SCAIL's full-context pose injection specifically preserves character appearance from the reference image while applying motion from the input video, maintaining cross-identity consistency.
Q: What parameters have the biggest impact on output quality?
A: Reference image quality and input video clarity are most critical. Resolution affects final detail, while seed controls variation. Prompt specificity guides anatomical understanding.
Popular Models
