Xingliu AI

Xingliu AI is a one-stop AI image generation platform from LiblibAI. Built on the in-house Star-3 large model, it is the official Chinese version of the globally popular design agent Lovart, with deep localization for Chinese users and a library of 100,000+ LoRA style models.

Core Capabilities

• Image Generation: Based on the Star-3 model, with fine image detail and zero AI look, supporting HD upscaling and smart image expansion
• Inpainting: Precisely modify parts of an image, fixing common AI problems such as hands
• Style Library: 100,000+ LoRA style models, with 200+ high-aesthetic styles added daily
• Multimodal Creation: One-stop generation of images, videos, 3D models, and audio
• Infinite Canvas: A professional editing environment with natural language interaction

Use Cases E-commerce product images, marketing materials, illustration, and short video assets. Especially suited to designers, e-commerce sellers, and content creators who need Chinese-language support and localized service.

Unique Advantages Compared with overseas products like Midjourney, Xingliu AI offers native Chinese semantic optimization and a Chinese-aesthetic style library, and its 100,000+ LoRA model library is industry-leading in scale — an ideal choice for users in China.

Editor's Recommendation Rated 4.3 on the App Store, with users generally praising its fast generation speed and good quality. Star-3 output approaches photorealistic quality, and the daily free allowance of 150 credits is enough to experience the core features.