This model is no longer supported.
Customize your input with more control.
Customize your input with more control.
Type # to reference inputs.
Hint: Drag and drop image files from your computer, images from web pages, paste from clipboard (Ctrl/Cmd+V), or provide a URL. Accepted file types: jpg, jpeg, png, webp, gif, avif, heic, heif
Customize your input with more control.
Your request will cost $0.3 per video.
Your request will cost $0.3 per video.
Hunyuan's SkyReels V1 delivers human-centric video generation from static images at $0.30 per video through fine-tuning HunyuanVideo on 10M+ film and television clips. This model emphasizes specialized human motion quality, focusing on facial expressions, body language, and natural movement patterns trained specifically on cinematic footage. Built for creators who need believable human subjects without the uncanny valley effect common in general video models.
Use Cases: Character Animation for Film | Social Media Content Creation | Marketing Video Production
At $0.30 per video generation, SkyReels V1 positions as a specialized alternative to general-purpose video models, trading broader scene capabilities for superior human motion fidelity.
| Metric | Result | Context |
|---|---|---|
| Output Format | MP4 video | Single image input to video output |
| Inference Steps | 1-50 configurable | Default 30 steps balances quality and speed |
| Cost per Video | $0.30 | 3.3 generations per $1.00 on fal |
| Aspect Ratios | 16:9, 9:16 | Optimized for social and widescreen formats |
| Guidance Scale | 1.0-20.0 range | Default 6.0 for balanced prompt adherence |
SkyReels V1 builds on HunyuanVideo's foundation with targeted fine-tuning on 10 million film and television clips, a dataset specifically curated for human subjects rather than general scenes. This training approach contrasts with broad video models that attempt universal scene generation by prioritizing natural human movement patterns, facial micro-expressions, and body language coherence.
What this means for you:
Cinematic Human Motion: Training on professional film footage produces movement quality that matches cinematography standards rather than synthetic motion patterns
Prompt-Driven Control: Text descriptions guide video generation with configurable guidance scale (1.0-20.0) and negative prompts to refine unwanted attributes
Flexible Aspect Ratios: Native support for 16:9 widescreen and 9:16 vertical formats eliminates post-production cropping for platform-specific content
Deterministic Generation: Optional seed parameter enables reproducible results for iterative refinement workflows
| Spec | Details |
|---|---|
| Architecture | HunyuanVideo (fine-tuned) |
| Input Formats | Single image URL (JPG, PNG, WebP, GIF, AVIF) + text prompt |
| Output Formats | MP4 video |
| Inference Steps | 1-50 configurable (default: 30) |
| License | Commercial use permitted |
API Documentation | Quickstart Guide | Enterprise Pricing
MuseTalk Image to Video – SkyReels V1 generates full-body human motion from single images with text prompt control. MuseTalk focuses specifically on audio-driven facial animation and lip-sync for talking head videos.
Kling Video v2.6 Pro Image to Video – SkyReels V1 trades general-purpose video generation for specialized human motion quality at $0.30 per video. Kling v2.6 Pro handles broader scene types with advanced prompt interpretation for multi-purpose video generation needs.