LM Studio Glossary
LM Studio is a free desktop application (macOS, Windows, Linux) for downloading and running open-weights large language models locally. The application provides a polished GUI on top of the open-source llama.cpp inference engine: a chat interface, a model-browser that pulls from Hugging Face, automatic detection of the user’s hardware capabilities (memory, GPU acceleration), an OpenAI-compatible local API server, and per-model parameter tuning (temperature, context length, prompt format).
For this Dictionary, LM Studio matters as the most polished consumer-facing local-LLM runner as of 2026. Where Ollama is command-line-first and developer-oriented, LM Studio offers a path into local AI inference for operators who are not comfortable on the command line. The application’s curated model picker and one-click hardware-appropriate model downloads remove most of the friction that historically kept local-LLM use a developer-only activity.
The Dictionary’s Sovereign Compute calculator treats LM Studio as the default recommendation for new operators entering the local-compute tier — with Ollama as the alternative for command-line-comfortable users. The operator’s own setup uses both, depending on whether the task at hand wants the GUI or the API.
See also
- Sovereign Compute
- Ollama — the command-line alternative
- Hugging Face — model source
- Hermes, Gemma, Qwen — common targets