lmm
A command-line local AI model manager for macOS and Linux that stores models once in the Hugging Face cache and exposes them to tools such as LM Studio, llama.cpp, Jan and ComfyUI through symlinks, avoiding duplicate downloads.
lmm manages local AI models across separate tools, such as LM Studio, Ollama, llama.cpp and ComfyUI, that would otherwise each keep their own copy of the same model, multiplying disk use. It uses the Hugging Face cache as a single canonical store and symlinks models into each tool's own directory for zero duplication and instant availability.
Features include interactive terminal search and installation of Hugging Face models in MLX, GGUF or safetensors format, adapters for LM Studio, llama.cpp, Jan, ComfyUI and text-generation-webui, adoption of models already present in the Hugging Face cache without re-downloading, a doctor command to validate integrity and a gc command to reclaim disk space, and shell completions for bash, zsh, fish, elvish and PowerShell.
Built with Rust, installable through Homebrew or cargo install; building from source requires Rust 1.87 or later. Apache-2.0 licensed.