One unified API for AI image generation across multiple models
Video-to-video editor for swapping characters, outfits and scenes while preserving camera motion
Stage
Not given
Not given
Founded
Not given
Not given
Based in
CN
Not given
Pricing model
Freemium
Paid
Pricing
Free: 3 images daily in web playground. Paid: GPT Image 2.6 Standard $0.0054/image at 1K resolution; GPT Image 2.6 Flare $0.0179/image at 1K, $0.0215 at 2K, $0.0269 at 4K; GPT Image 2.6 Sunburst $0.0215/image at 1K, $0.0269 at 2K, $0.0322 at 4K; GPT Image 2.5 Standard $0.0054/image at 1K, $0.0143 at
Not given
Key features
Support for multiple AI models including GPT Image 2.6, GPT Image 2.5, Gemini (Banana) and Seedream 5.0 Pro
Text-to-image generation from custom prompts
Image-to-image generation using reference images
Adjustable image resolution up to 4K
Multiple aspect ratio options for generated images
Credit-based billing with no failed charge for unsuccessful generations
Free daily allowance of 3 images without account signup
Support for up to 12 reference images per request
Consistent API parameters across all supported models
Swap characters, outfits, vehicles and objects using reference images
Motion transfer to new performers or settings while preserving movement
Three AI models: Standard (Kling O3 or FLUX.3), Lite (Genjutsu Lite), Pro (Higgsfield Genjutsu)
See exact credit cost before rendering
Failed renders automatically refunded
Render clips 4–30 seconds at 480p or 720p
What makes it different
Unified API access to multiple top AI image generation models through a single interface
No account or signup required to try 3 free images daily
Pay only for successful image generation with automatic credit refunds on failures
Supports both text prompts and reference images for flexible generation workflows
Latest high-quality models including GPT Image 2.6 and Seedream 5.0 Pro
No subscription: pay only for renders with one-time packs or optional monthly plans
Access official Higgsfield Genjutsu without subscription through the API
See exact cost before generating and get refunds for failed renders
Simple interface: upload, describe, render with no timeline or keyframes
Private renders never published to public galleries