techjarves/Uncensored-Local-Studio

Uncensored local AI studio for Windows, Linux, and macOS. Zero-setup GUI for Image Generation, GGUF LLMs, Text to Speech & Speech to Text

What it solves

Uncensored AI Studio provides a unified, completely offline desktop interface for running multiple types of AI models. It eliminates the need for cloud subscriptions, API keys, and internet connectivity, ensuring total privacy and removing censorship from AI interactions.

How it works

The studio acts as a high-performance GUI that manages several specialized local backends. It uses a workspace-based architecture to prevent system memory exhaustion by switching between different engines:

  • Image Generation: Powered by stable-diffusion.cpp for Stable Diffusion models.
  • Text Chat: Powered by llama.cpp for GGUF-format language models.
  • Speech-to-Text: Powered by whisper.cpp for audio transcription.
  • Text-to-Speech: Powered by kokoro-js using the Kokoro-82M ONNX model.

It automatically detects hardware to apply the best acceleration (CUDA for Nvidia, ROCm for AMD, Metal for macOS, or OpenVINO for Intel NPUs).

Who it’s for

Users who want a "zero-configuration" experience to run diverse AI modalities (text, image, audio) locally on Windows, Linux, or macOS without relying on external servers or cloud providers.

Highlights

  • Zero-Install Portability: Includes a self-contained runtime (Node.js and GPU backends) that doesn't change global system environments.
  • Hardware-Accelerated: Native support for Nvidia, AMD, Intel, and Apple Silicon GPUs/NPUs.
  • Integrated Model Manager: Allows users to download weights directly from Hugging Face or import local files via drag-and-drop.
  • Real-time Monitoring: Built-in tracking for CPU, RAM, GPU, and VRAM usage within the web UI.