AI Potluck
Contribute
Back to Gap Map Model components / Image, video, 3D & music generation

stable-diffusion.cpp

leejet
open source / Overall score: 2.6

stable-diffusion.cpp is a C and C++ inference engine for diffusion image and video models, built on the ggml tensor library and run from a command line or an embedded web interface. It reads safetensors and GGUF weights and targets CPUs and CUDA, Vulkan, Metal, OpenCL and SYCL devices. Supported models range from the Stable Diffusion line through FLUX and Qwen-Image to the Wan, LTX and HunyuanVideo video models.

The only Python package for it on PyPI is a third-party binding from a different author and repository, so it is not declared here.

Openness

5 high confidence
5.0
license
MIT(OSI)
source
public(leejet/stable-diffusion.cpp)
core features withheld
no — no ee, enterprise, commercial or proprietary path

stable-diffusion.cpp is MIT-licensed with its full source public. It is maintained as an independent project and no paid or hosted edition is offered beside it.

Adoption

2 low confidence
2.0

The project publishes prebuilt binaries on its release page rather than through a package registry, so GitHub stars are the only comparable signal. A star is not a use.

Capability

3 high confidence
3.0

stable-diffusion.cpp runs both image and video generation models across many families, which few engines here do. It stops short of ComfyUI because it runs no audio or 3D models and composes no workflows.

Verified 2026-09-26