llmfit
Terminal tool that matches local LLMs to your CPU, RAM, and GPU
What it is
llmfit detects your CPU, system memory, GPUs, and VRAM, then rates hundreds of open models and quantizations on how well they fit that machine, how fast and how well they would run, and how much context they allow. It offers an interactive TUI, a scriptable CLI with JSON output, a web dashboard, and a REST API, and can hand models off to Ollama, llama.cpp, MLX, or LM Studio. People running models locally use it to pick a size and quantization before downloading, and can benchmark real tokens per second and share the results.