LM Studio Glossary
LM Studio is a desktop application and local server for downloading and running open-weight language models on macOS, Windows, and Linux. It provides chat and model-discovery interfaces, hardware-aware loading controls, and OpenAI-compatible local API endpoints. It supports GGUF models through llama.cpp and, on Apple silicon, MLX models.
For this Dictionary, LM Studio matters as a polished consumer-facing local-model runner. Where Ollama is command-line-first, LM Studio offers a graphical path into local inference. Model discovery still requires judgment: a downloadable model may be too large for the machine, governed by a restrictive licence, or unsuitable for the task.
The Dictionary’s Sovereign Compute framework treats LM Studio as one accessible starting point, with Ollama as an alternative for command-line-comfortable users. The operator’s own setup uses both, depending on whether the task wants a GUI or a service-oriented workflow.
See also
- Sovereign Compute
- Ollama — the command-line alternative
- Hugging Face — a model source
- Hermes, Gemma, Qwen — common targets
Sources
- LM Studio, Welcome to LM Studio Docs: https://lmstudio.ai/docs/app
- LM Studio, Offline operation: https://lmstudio.ai/docs/app/offline