LM Studio

As explained in Domain-Specific Small Language Models

A local model-serving application and desktop environment that lets users run, chat with, and inspect models offline on their own machine.

LM Studio is a desktop application for downloading, managing, chatting with, and serving language models locally. After models are downloaded, it supports entirely offline inference on macOS, Windows, and Linux. It is free but not open source, while providing open source utilities such as Python and JavaScript SDKs and the Apple MLX engine. section 12.3

LM Studio combines several operational paths:

• Chat view supports conversation, chat history, system prompts, temperature, sampling parameters, and document attachments. section 12.3
• Document attachments accept .docx, .pdf, and .txt files; shorter documents enter the conversation directly, while longer documents use chunking and retrieval-augmented generation. section 12.3
• Developer view and the local server support client applications through a REST API with OpenAI-like behavior. section 12.3
• The companion lms CLI loads and unloads models, starts or stops the local API server, and inspects raw model input and output. section 12.3

Typical local workflow

Download a GGUF or Apple MLX model from the LM Studio catalog or Hugging Face Hub, select it in the chat view, attach supporting documents when needed, and run conversations offline. Developers can instead start the local server and connect an application through the REST API or Python SDK. section 12.3 section 12.7

Figure 12.5 LM Studio’s chat view

The chat-view figure helps locate model selection, conversation history, advanced generation settings, and document attachment controls. section 12.3

How Domain-Specific Small Language Models covers it

How deep it goes

Explains desktop chat, document retrieval, model formats, CLI operations, REST serving, and speculative decoding without detailing server internals.

Examples

Uses Llama 3.2 3B Instruct, attached documents, the LM Studio catalog, and draft-target speculative decoding models.

What it assumes

Assumes access to a compatible computer and a downloaded model; introduces chat and developer workflows for varied technical skill levels.

What you can do after

Operate/configure local models through the desktop interface, CLI, REST API, or Python SDK.

How it connects

Connects local inference with GGUF, Apple MLX, retrieval-augmented generation, REST APIs, Python SDKs, and speculative decoding.

liveBookmanning.com

Read Domain-Specific Small Language Models on liveBook Get the book at manning.com

Also explained in

• 

  liveBookmanning.com
  Data-Driven CybersecurityLM Studio is a desktop application for downloading, running, and interacting with language models through a graphical interface.
• 

  liveBookmanning.com
  AI Agents in ActionLM Studio is a desktop application for running large language models locally and exposing them through a server interface.

Compare how 3 books explain LM Studio →

Related concepts

• language modeluses · A language model is a computational model that represents patterns in linguistic tokens to estimate sequence probabilities and predict or generate language from context.
• Ollamaalternative to · Ollama is software for downloading, running, and serving language models locally.
• user interfaceuses · A user interface is the part of a system through which people provide input, receive information, and interact with its functionality.
• GGUFuses · GGUF is a binary file format for storing machine-learning model tensors and associated metadata, designed to support efficient model loading and inference.
• OpenAI APIalternative to · The OpenAI API is a programming interface that allows software applications to send requests to OpenAI models and services and receive responses.

This explainer was generated by Manning's knowledge graph from the text of Domain-Specific Small Language Models. The section links point to the parts of the book it draws on.