行者AI美术

1. Judgment: Instead of another Midjourney, this is a production-oriented platform wrapping Stable Diffusion workflows, focused on controllability rather than random generation. It moves local setup, parameter tuning and model training to the cloud for consistent, high-precision pipelines. 2. Function: It includes ‘Tu Shuashua’ (online drawing with text/img2img, inpainting, upscaling, ControlNet, pose editing, etc.) and ‘Xingzhe Danlu’ (free LoRA trainer to create custom style/character models from uploaded images). CEO claims a 3–5× efficiency boost for game concept artists when generating concept art and three-view drawings. 3. Threshold & cost: Tu Shuashua is free but capped at 400 lifetime images (per official info); Danlu training is free. No public pricing for advanced commercial use or API. Suited for game studios and indie developers who need rapid iteration with controllable output and custom model training; not for casual users wanting unlimited free generation or standalone artistic exploration without pipeline integration. 4.