LM Studio

LM Studio is a desktop application for downloading, running, and interacting with large language models locally, akaLocal AIon your own hardware, rather than through a cloud API. Unlike more terminal-driven tools, LM Studio is built around a graphical interface — browsing and downloading models, chatting with them directly in-app, and comparing multiple models side by side without touching a config file or command line.

It exposes adjustable parameters like context length, temperature, and system prompts through its interface, and runs a local API server compatible with the same format many cloud providers use — meaning it can often be dropped in as a drop-in local replacement for a cloud API in other tools that expect that format.

Like other local-inference tools, LM Studio supports quantized models sized to run on modest consumer hardware, making it practical to experiment with open-weight models entirely offline, with no data leaving the machine.

Role in this vault

LM Studio offers a more visual, hands-on way to experiment with local AI models that could eventually support this vault — useful for testing and comparing models before settling on one to run continuously via Ollama for any retrieval-augmented search built on top of these notes.