Longcat Image

We built Longcat Image because rendering readable text inside AI images was still unreliable. The LongCat-Image model finally nails Chinese and English typography, so we wrapped it in a fast editor anyone can use. Feedback welcome!

Really like that this is built on Meituan's open-sourced LongCat-Image 6B model instead of a closed one — 6B is a pretty small footprint compared to most competing image models, and the Chinese text rendering samples on the page look noticeably cleaner than what I've seen from bigger closed models. Question on the multi-round editing: how does it preserve facial identity/subject consistency across rounds — is that built into the base model's training, or is there a separate identity-locking step? A before/after strip of the same subject through 3-4 edit rounds on the landing page would make that claim much easier to trust at a glance.

Going with Meituan's open 6B model instead of a bigger proprietary one is an interesting bet, especially given the note that it renders Chinese text more cleanly than some larger competitors text rendering is usually where image models fall apart. Is the compactness mainly a cost/speed tradeoff for you, or did you find the smaller model was actually competitive on quality too?

interesting but i want more free option

Great combination of high-quality image generation and precise text rendering. The natural-language editing makes creating polished marketing and e-commerce visuals much faster. 🔗 https://pixprep.online/

This looks great, I'll give it a try

How is it different from other platforms ?

It is a good image generator for local and private image generation (If u use the local model of theirs). Even as cloud option you can use it for normal use cases.

I really like the style and the homepage; the way you create the images is almost addictive.