The outcome

Select a model configuration that provides acceptable quality, speed, and memory use on available hardware.

LM Studio is a desktop and headless runtime for downloading, evaluating, and serving local language models. Running supported open models locally, comparing model configurations, chatting with private local data, and providing local application APIs. This guide narrows that broad capability into one repeatable outcome, with checkpoints that keep the source material and your judgment in the loop.

Before you begin

Set the boundary before the tool starts.

Choose one real task, identify who will use the result, and decide what evidence or test will make the result acceptable. Gather only the source material needed for that task. If the work contains confidential, personal, regulated, or client-owned information, confirm that the platform and account are approved before sharing it.

Troiana principle

AI should make the work easier to inspect. If the workflow removes the source, the owner, or the review step, redesign the workflow.

Step by step

A workflow you can repeat.

  1. 01

    Record the task, languages, context size, license needs, privacy boundary, and the machine's RAM, VRAM, processor, and free storage.

  2. 02

    Search LM Studio for compatible GGUF or MLX candidates and inspect their publisher, model card, license, parameter size, and quantization.

  3. 03

    Download a realistic size, load it with a conservative context and hardware-offload setting, and confirm the app remains stable.

  4. 04

    Run the same representative, factual, and adversarial prompts while recording quality, latency, memory use, and failure patterns.

  5. 05

    Save the chosen model identifier and settings, verify offline operation if required, and retain a smaller fallback for constrained machines.

Working standard

What good use looks like.

  • Choose models for the actual hardware.
  • Evaluate several task-shaped prompts.
  • Record model and quantization exactly.

Local inference does not make model output trustworthy, and the API server has no authentication by default. Review model licenses and cards, keep the server on localhost unless necessary, require a token before network exposure, and tightly control CORS and MCP access.

Official references

Check the current product documentation.

Features, plan limits, availability, and data controls change. These official pages are the starting points used for this collection.