Configure
Shape model and runtime settings through the browser. Leverage multiple presets per model, assigned to arbitrary multiple llama.cpp versions.
GGUF · llama.cpp · browser-native
Inference,
under control.
VisLlama is web-based browser administration for GGUF models served through llama.cpp. Configure, launch, route and monitor inference from one focused control surface.
Open source · GNU GPLv3
Launch
Bring llama.cpp-backed GGUF inference online. Mix and match any version of llama.cpp.
Route
Direct inference traffic across your served models. Multiple versions of llama.cpp unified into one OpenAI-compatible API endpoint.
Monitor
Keep GPU/system, model activity and inference behavior in view, build stats of model performance over time and with different llama.cpp versions.