# LocalLLM

A lightweight Tauri desktop control panel for running llama.cpp locally, letting a user discover GGUF models, configure and preview the launch command for llama-server, and save per-model presets.

LocalLLM is a desktop application for managing a local llama.cpp installation without hand-editing commands. It discovers llama-server, llama-cli, and Hugging Face download tools already on the system path, can install prebuilt llama.cpp release assets for Windows and Ubuntu Linux from inside the app, and scans a chosen model directory for GGUF files, or lets a user add files from elsewhere.

It configures runtime, KV cache, GPU and memory, sampling, speculative decoding, multimodal, and tool-calling options, previews the exact llama-server command before launch, and estimates VRAM use from GGUF metadata. Server presets are saved per model, the server can run hidden in the background or in a visible terminal, and the app also manages versioned agent profiles with their own workspace, instructions, and permissions, recovering automatically from an interrupted configuration write.

Website: https://github.com/TRI-Tech-Revolution-Intelligence/LocalLLM