Starter: $5.99 for 1 credit; Creator: $14.99 for 3 credits; Studio: $39.99 for 10 credits. One-time purchase, no subscription. Standard creation uses 1 credit, enhanced video mode uses 2 credits.
Key features
Pyramid-JiT (P-JiT) architecture with 32×32 token reduction
4× more aggressive token reduction than Linum v2
Converges in 11× fewer samples than Linum v2
2B-parameter model with frozen Qwen text encoder
Pixel-space diffusion with multi-resolution readouts
PixelREPA alignment with DINOv3-L
Curriculum learning over noise schedule
AI-generated costume transformations with preset options including wizard, witch, vampire, and astronaut
Three captioned images and 9-second silent MP4 export at 720×1280 resolution
Edit captions after generation without requiring regeneration
Private account-based storage for uploads and saved results
Mobile browser support with optional image-only download
Shared credit balance across three templates: I'm Sorry I Lied, Dime, and Ozil
Failed generations automatically return credits
Free fictional samples available to watch and download
What makes it different
Low-cost training: full 2B model trained for just $21K in compute
Controls baked into model training rather than post-hoc fine-tuning
Pixel-space compression without separate VAE bottleneck
Apache 2.0 open release with full model weights and code
Single-photo creation workflow requiring only one original image
No subscription required; credits never expire and are purchased as one-time packs
Integrated platform handling both AI generation and slideshow creation in one tool
Silent MP4 downloads for flexible music and audio customization