DEV.co
AI Frameworks · lmstudio-ai

lms

LM Studio CLI is a command-line interface for managing and interacting with LM Studio, a desktop application for running large language models locally. It provides tools to start/stop servers, load/unload models, check system status, and stream logs.

Source: GitHub — github.com/lmstudio-ai/lms
5k
GitHub stars
420
Forks
TypeScript
Primary language
MIT
License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

FieldValue
Repositorylmstudio-ai/lms
Ownerlmstudio-ai
Primary languageTypeScript
LicenseMIT — OSI-approved
Stars5k
Forks420
Open issues309
Latest releaseUnknown
Last updated2026-07-06
Sourcehttps://github.com/lmstudio-ai/lms

What lms is

TypeScript-based CLI built on the lmstudio.js SDK, distributed as part of LM Studio 0.2.22+. It wraps local API server management and model lifecycle operations via command-line subcommands, with JSON output support for automation.

Quickstart

Get the lms source

Clone the repository and explore it locally.

terminalbash
git clone https://github.com/lmstudio-ai/lms.gitcd lms# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Local LLM Server Orchestration

Automate starting/stopping LM Studio's inference server and manage loaded models without GUI interaction—useful for headless deployments or CI/CD workflows.

Development & Testing Workflows

Developers building on lmstudio.js SDK can use CLI to quickly inspect downloaded models, check server status, and load models for testing locally.

Script-Based Model Management

Batch operations on models (listing, loading, unloading) via JSON output enable integration into shell scripts, Python automation, or DevOps pipelines.

Implementation considerations

  • Requires LM Studio 0.2.22 or newer installed on the target machine; CLI is not independently distributable.
  • JSON output mode (`--json` flags) enables programmatic parsing; shell integration should validate JSON structure before consuming.
  • Model loading with `-y` flag bypasses confirmation prompts; automation scripts should handle GPU/memory constraints and load failures gracefully.
  • Log streaming via `lms log stream` may buffer large outputs; consider redirecting to file or log aggregation for long-running inference jobs.
  • CLI runs against the LM Studio daemon; ensure server is responsive before issuing CLI commands to avoid timeout errors.

When to avoid it — and what to weigh

  • Distributed or Cloud-Native Inference — LM Studio is designed for local/desktop execution. If you require multi-node, GPU-cluster, or cloud-hosted LLM serving, use alternatives like vLLM, Ray Serve, or cloud providers.
  • Production-Grade Security Requirements — CLI documentation does not address authentication, authorization, or audit logging. Not suitable for multi-tenant or regulated environments without significant hardening.
  • Frequent Zero-Downtime Updates — CLI is bundled with LM Studio desktop application; CLI updates require application updates and may require service restarts.
  • Cross-Platform Container Deployment at Scale — LM Studio CLI is tightly coupled to the desktop app. Containerization and orchestration patterns are not documented or standard.

License & commercial use

Licensed under MIT (MIT License), a permissive OSI-approved open-source license. Permits use, modification, and distribution in commercial and proprietary projects, with attribution required.

MIT license permits commercial use. However, the CLI is distributed as part of LM Studio (a separate application). Review LM Studio's own license terms and usage restrictions for production deployments. No warranty or liability waiver is provided by the MIT license alone.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

SignalAssessment
MaintenanceActive
DocumentationAdequate
License clarityClear
Deployment complexityModerate
DEV.co fitPossible
Assessment confidenceHigh
Security considerations

CLI assumes local, trusted execution only. No authentication, encryption, or network isolation documented. Model loading and server management have no rate-limiting or audit trail. Suitable only for personal/development use; not recommended for multi-user, internet-facing, or regulated environments.

Alternatives to consider

vLLM

Production-grade inference server with distributed serving, advanced batching, and OpenAI-compatible API. Requires separate deployment but scales to multi-GPU/multi-node.

Ollama CLI

Lightweight local LLM CLI alternative with model download/management similar to `lms`, broader model support, and simpler API. No monorepo dependency.

Ray Serve + Ray LLM

Distributed inference framework with scaling, fault tolerance, and cloud-native deployment. Steeper learning curve but suitable for production multi-model serving.

Software development agency

Build on lms with DEV.co software developers

LM Studio CLI is ideal for development and testing. If you need production-grade, distributed, or cloud-hosted LLM serving, we can help you evaluate alternatives or design a custom solution.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

Related on DEV.co

Explore the category and the services that help you build with it.

lms FAQ

Can I use LM Studio CLI without the desktop application?
No. CLI is bundled with LM Studio 0.2.22+ and requires the LM Studio daemon to be running. It cannot operate independently.
Is the CLI suitable for production inference?
Not for multi-user or internet-facing production. It is designed for local, trusted development/testing. Consider vLLM or Ollama for production deployments.
How do I automate model loading in CI/CD?
Use `lms load <model-path> -y` to load without confirmation, and capture JSON output with `--json` flags for scripting. Ensure LM Studio daemon is started before CLI calls.
What is the difference between `lms ls` and `lms ps`?
`lms ls` lists all downloaded models on disk; `lms ps` lists only currently loaded models in memory available for inference.

Work with a software development agency

DEV.co helps companies turn open-source tools like lms into production software. Our software development services cover the full lifecycle — architecture, web development, integration, and maintenance — delivered by software developers and web developers who ship. Engage our software development agency to implement or customize it for your ai frameworks stack.

Ready to integrate local LLM inference into your workflow?

LM Studio CLI is ideal for development and testing. If you need production-grade, distributed, or cloud-hosted LLM serving, we can help you evaluate alternatives or design a custom solution.