alexsjones/llmfit

llmfit is a terminal tool that helps you figure out which AI language models (LLMs, large language models, the kind that power chatbots and AI assistants) will actually run well on your specific computer hardware. The core problem it solves: there are hundreds of LLM models available for local use, but each one has different memory and compute requirements, and downloading one only to find your machine can't run it well is frustrating and time-consuming. You run llmfit in your terminal, and it automatically detects your computer's RAM, CPU, GPU, and available VRAM (graphics card memory). It then scores each model across dimensions like quality, speed, and how well it fits your hardware, and shows you a sortable, filterable list so you can find the best match. You can search by name, filter by whether a model will run comfortably or just barely, and browse community benchmark data showing real performance numbers from other users with similar hardware. It also includes a download manager so you can grab a chosen model directly from the interface. llmfit works with local runtime backends like Ollama, llama.cpp, MLX, LM Studio, and Docker Model Runner. You would use it when you want to run AI models locally, for privacy, cost, or offline use, and you want guidance on which one to choose before committing to a download. The tech stack is Rust.