Imagine anything, create instantly

Sign in, pick a mode (Text→Image), type a description of what you want, choose aspect ratio, click Generate. Generation takes ~20-60s depending on resolution and steps.

Pick a persona card (Seller/Designer/Illustrator/Hobbyist) to auto-apply the right ratio, prompt template, negative prompt and internal quality suffix. Click prompt chips to append professional terms to your prompt.

Every user gets 10 free generations per day. After that, each image costs 1 credit. $9.9 buys 1000 credits via Paddle — credits never expire and are only charged on successful submissions.

Denoise (img2img) controls how much the image changes: 0.2-0.4 keeps composition, 0.6+ restyles. Steps: 20-30 for quality, 4-8 for quick drafts. CFG: higher = stricter prompt following. Same seed + same params = reproducible image.

Upload any image and pick a tool mode — no prompt needed. Remove BG outputs transparent cutouts; Restore fixes scratches on old photos; Face Fix repairs blurry faces; Deblur recovers pixelated/compressed images; Upscale enlarges 4x.

Yes. Registration is free and each user gets 30 image generations per day at no cost.

Text-to-image generation, image-to-image editing, inpainting/local redraw, and 4x upscaling, with persona presets for e-commerce sellers, designers, illustrators and hobbyists.

Yes, the interface is fully bilingual in English and Chinese, and the model understands Chinese prompts.

No. All generated images are watermark-free and can be downloaded in full resolution.