| Index | index by Group | index by Distribution | index by Vendor | index by creation date | index by Name | Mirrors | Help | Search |
| Name: stable-diffusion.cpp-server | Distribution: OpenMandriva Lx |
| Version: 0.820 | Vendor: OpenMandriva |
| Release: 1 | Build date: Sun Aug 16 16:28:54 2026 |
| Group: Servers | Build host: ryzen9-5.openmandriva.org |
| Size: 0 | Source RPM: stable-diffusion.cpp-0.820-1.src.rpm |
| Packager: bero <bero@lindev.ch> | |
| Url: https://github.com/leejet/stable-diffusion.cpp | |
| Summary: HTTP inference server for stable-diffusion.cpp | |
HTTP server (sd-server) for stable-diffusion.cpp. OpenAI-style
(/v1/images/generations), Automatic1111 (/sdapi/v1/txt2img) and
native /sdcpp/v1 routes. Built without the pnpm frontend (ABF has
no network); the JSON API still works.
Config: /etc/sysconfig/sd-server
Unit: stable-diffusion.service (default 127.0.0.1:8082, not
upstream 1234, so it can run next to llama-server and
whisper.service)
systemctl enable --now stable-diffusion.service
If MODEL / DIFFUSION_MODEL are unset, the unit picks the first
readable .gguf/.safetensors/.ckpt/.pt under MODEL_DIRS
(/srv/ai then /usr/share/stable-diffusion.cpp/models).
Split pipelines (Flux, Z-Image, Qwen-Image, Wan) need
DIFFUSION_MODEL plus VAE / LLM / CLIP_L / T5XXL — do not rely
on the auto-pick for those.
There is no built-in API key. Do not bind HOST to a public
address without a reverse proxy.
List ggml devices: sd-server --list-devices
To generate (after a model is loaded):
curl http://127.0.0.1:8082/v1/images/generations \
-H "Content-Type: application/json" \
-d '{"prompt":"a cat","size":"512x512"}'
curl http://127.0.0.1:8082/sdapi/v1/txt2img \
-H "Content-Type: application/json" \
-d '{"prompt":"a cat","width":512,"height":512}'
MIT
/etc/sysconfig/sd-server /usr/bin/sd-server /usr/lib/systemd/system/stable-diffusion.service /usr/libexec/stable-diffusion.cpp/sd-server-start
Generated by rpm2html 1.8.1
Fabrice Bellet, Tue Aug 18 22:00:29 2026