AI Song Generation on Mac Apple Silicon, with Full Style Control - Generate complete songs with lyrics, vocals, and instrumental tracks using Tencent AI Lab's SongGeneration (LeVo) model.
Radar
New projects people are discovering or following across Pinokio.
Fast and High-Quality Zero-Shot voice clone Text-to-Speech with Flow Matching
Local generative video, image, and character training on Apple Silicon. Train face + voice LoRAs in-app. Q8 HQ for character clips. MLX native — no cloud, no API key.
Local-first AI video editor with optional native MiniMax H3 generation on Apple Silicon.
[NVIDIA ONLY] a state-of-the-art open-source model for fast feedforward 3D mesh reconstruction from a single image, from Stability AI. https://huggingface.co/spaces/stabilityai/stable-fast-3d
Upload the picture of an image, and generate images with that image style. Instant generation with no LoRA required https://huggingface.co/spaces/InstantX/InstantStyle
Run a whole folder of images through an existing FaceFusion 3.6.1 install in one click. Pick an input folder, an output folder, a source face and the processors, then press Run Batch.
FLUX.3 Image Klein studio (5.6B, 16 steps, FLUX Kommunity License), Fooocus-style, 100% local: txt2img, multi-reference editing in the SAME pipeline (<ref_image_0>, <ref_image_1>..., no second model), inpaint/outpaint, ESRGAN+refine upscale, LoRA, styles, Describe/Improve & Vision Mix (Ollama), Remove BG, Face Swap. 18.8 GB of VRAM in bf16 for everything (fp8r / nvfp4 builds for smaller cards). The repos are gated on Hugging Face: accept the licence and set a token. Fork of crispz-klein. https://github.com/mikecastrodemaria/crispz-flux3
A universal speech enhancement model for diverse degradations, sampling rates, and languages made by Nvidia, using Nemotron RE-USE's Hugginspace Space clone for local inference. Compatible with Windows and CUDA 12.8 GPUs
