Repolode
Submit
rupeshs

rupeshs/fastsdcpu

A Stable Diffusion image generator optimized to run on CPUs and Intel AI PCs, with desktop GUI, web UI, CLI and API interfaces.

View on GitHub
Stars
2,188
Forks
218
Open issues
64
Last commit
4 days ago

What is it?

FastSD CPU is a Stable Diffusion image generator designed to produce images quickly without a dedicated GPU. It builds on Latent Consistency Models and Adversarial Diffusion Distillation to reduce generation to a few steps, and it can use PyTorch or Intel OpenVINO as the inference backend.

The project offers three ways to use it: a Qt desktop app for basic text-to-image generation, a WebUI with advanced options such as LoRA and ControlNet, and a command-line interface. It also exposes a REST API, an MCP server for AI clients such as Claude Desktop, and a Docker setup. Supported platforms include Windows, Linux, macOS, Android via Termux, and Raspberry Pi 4.

FastSD CPU supports SD 1.5, SDXL, SD Turbo, SDXL-Lightning, Hyper-SD, FLUX.1 schnell and FLUX.2 klein models, along with image-to-image and image editing workflows. OpenVINO builds can target CPU, integrated GPU and Intel NPU.

Key features

  • 01Desktop GUI, WebUI and command-line interfaces
  • 02OpenVINO backend for CPU, integrated GPU and Intel NPU
  • 03LoRA, ControlNet and multi-LoRA support in the WebUI
  • 04REST API and MCP server for integration with other tools
  • 05Supports GGUF quantized Flux 2 Klein models and image editing presets
  • 06Runs on Windows, Linux, macOS, Android (Termux) and Raspberry Pi 4

Use cases

  • Generating images on a laptop or desktop without a dedicated GPU
  • Running Stable Diffusion on Intel AI PCs using CPU, GPU or NPU
  • Connecting local image generation to Open WebUI, Claude Desktop or ComfyUI
  • Restoring or colorizing old photos with the built-in editing presets

Things to know

  • Some bundled models, such as the SD Turbo models, are intended for research purposes only.
  • The FLUX.1-schnell OpenVINO workflow needs around 30GB of system RAM for 512x512 images.
  • Real-time text to image and the SDXS-512-0.9 workflow are marked experimental.

In collections

MCP serversCLI & terminal toolsAI image generators & Stable Diffusion tools

Topics

#stable-diffusion#text-to-image#openvino#cpu-inference#lcm#image-generation#mcp#aipc#api#cli

FAQ

What is fastsdcpu?
A Stable Diffusion image generator optimized to run on CPUs and Intel AI PCs, with desktop GUI, web UI, CLI and API interfaces. FastSD CPU is a Stable Diffusion image generator designed to produce images quickly without a dedicated GPU. It builds on Latent Consistency Models and Adversarial Diffusion Distillation to reduce generation to a few steps, and it can use PyTorch or Intel OpenVINO as the inference backend.
Is fastsdcpu free to use? What license does it have?
fastsdcpu is open source under the MIT license. Check the license terms before using it in commercial products.
Is fastsdcpu still maintained?
Actively maintained. The last commit to fastsdcpu was on Oct 18, 2026.
What can I use fastsdcpu for?
Generating images on a laptop or desktop without a dedicated GPU Running Stable Diffusion on Intel AI PCs using CPU, GPU or NPU Connecting local image generation to Open WebUI, Claude Desktop or ComfyUI Restoring or colorizing old photos with the built-in editing presets
Where can I get fastsdcpu?
The source code and installation instructions are on GitHub: https://github.com/rupeshs/fastsdcpu

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects