Local image-to-3D model for Apple Silicon and iPhone
A port of Hunyuan3D-Paint and Hunyuan3D-Shape to Swift MLX and Python MLX, enabling image-to-3D generation on Apple Silicon Macs and iPhones with low memory usage. It solves the problem of running 3D generation locally without cloud dependencies or high-end GPUs.
Entities
Related
Xiaomi quietly uploads MiMo-V2.5-DFlash weights to Hugging Face
Xiaomi has uploaded the official DFlash weights for MiMo-V2.5-DFlash to Hugging Face. The model has 300B+ parameters and runs at 8-10 tokens/second on 2x24GB cards with offloading; DFlash could double that speed. The MTP head was shared but does not work yet.
MCP server to generate App Store screenshot mockups from raw screenshots
An open-source MCP server that takes raw app screenshots and produces App Store-ready preview images. It uses Claude to analyze app colors, propose themes and captions, and renders framed, captioned mockups (1284x2778) ready for upload to App Store Connect. Built for developers who need to create App Store preview images without manual design work.
MCP server that generates App Store mockups from raw screenshots
An MCP server that takes raw app screenshots and generates framed, captioned preview images ready for App Store Connect. It uses Claude/Cursor to analyze app colors, propose themes and captions, and renders 1284x2778 images with procedural iPhone frames via Pillow.
AI photo editor with text prompt editing
An AI-powered photo editor that lets users perform professional-grade image editing using text prompts. It uses LLMs to interpret natural language instructions and apply edits accordingly, solving the problem of complex manual editing for casual users and professionals alike.
Hugging Face introduces Kernels repository type with redesigned CLIs and security
Hugging Face announced a new 'kernel' repository type on the Hub, along with a major redesign of the project including improved security, revamped CLIs, expanded framework/backend coverage, and a foundation for agentic kernel development. The updates aim to better serve users with compute-specific needs.