Self-growing tag-dictionary RAG for any device. Zero dependencies, no GPU, no vector DB. pip install sandclaw-memory

A real-time digital human technology based on multimodal feature matching. GPU-free and CPU-native, it enables low-latency driving with minimal computational overhead. This project offers a lightweight, easy-to-integrate solution for resource-constrained, real-time interaction scenarios.

Free Google Colab notebook for running AUTOMATIC1111's Stable Diffusion WebUI. One-click setup with pre-configured settings, ADetailer for fixing hands/faces, and Cloudflare tunnel. No local GPU required — generate AI images in ~6 minutes.