llmfit
An open-source Rust CLI that detects local hardware and ranks runnable LLMs/providers, helping local AI developers quickly produce model selection and deployment guidance.
Tool overview
If your main question is “which local LLMs can this machine actually run?”, llmfit looks worth trying first. If you need model training, hosting, a unified inference gateway, or a chat UI, it is not that kind of product. A better analogy is a local LLM compatibility advisor and model-picking assistant, not a runtime like Ollama, not a fine-tuning stack, and not a cloud API aggregation layer.
The practical role is reasonably well supported by the evidence. The official GitHub repo explicitly says it helps find what runs on your hardware across hundreds of models and providers, and several X posts repeat that it checks things like VRAM, storage, and OS before surfacing options. That is enough to support the claim that it can generate a first-pass shortlist. But heat proof and usability proof are different: GitHub stars around 29k, Trending mentions, and repost-heavy social posts mostly show attention, while stronger support for actual capability comes from the repo itself and a few installation/use examples. Public hands-on benchmarks still look limited.