Onion AI

Onion AI is a generative image tool. A Chinese language, browser based AI image tool that generates pictures from a text prompt and lets you keep refining or editing them through continued conversation, including edits to uploaded photos and AI poster design.

Onion AI (小洋葱AI) is a browser based image generation and editing service built around a single chat style input. Users type one sentence describing the scene they want, pick an aspect ratio, and the platform renders a picture. From there the conversation continues: each follow up message revises the existing image instead of starting over, so adjustments accumulate in the same thread. The homepage offers sample prompts such as an orange cat floating in space in a Pixar style, a latte and croissant by a cafe window in soft film tones, and a rainy neon cyberpunk street, which shows the range of styles the model is expected to handle. Beyond text to image, the tool accepts uploaded photos and lets the AI modify them. The product listing groups these abilities under text to image, AI retouching, style transfer, smart photo repair and poster design, all aimed at people who need finished visuals without design software. Aspect ratio presets cover square, portrait 3:4, landscape 4:3, vertical full screen 9:16 and widescreen 16:9, with a smart ratio option that picks a format automatically. That makes it practical for social posts, vertical video covers, banners and print style posters. The audience is mostly casual and semi professional: small business owners making promotional images, social media creators who need variations quickly, and hobbyists experimenting with prompts. Because the main interaction is conversational captioning rather than layer based editing, the learning curve is short, and nothing about the interface assumes prior design training. The account area keeps a history of recent generations and shows remaining free quota and paid credits. Compared with prompt only generators, the differentiator is the edit by conversation loop, which suits iterative refinement more than one shot rendering. Compared with full design suites, it trades control for speed. The site is Chinese language first and the interface is web only, so teams needing API access, brand asset libraries or collaboration features will likely pair it with other tools.