LM Studio
Run downloaded language models and document chats locally. Optional Bionic cloud inference has its own plans and usage credits.
- Starting price
- $0 local software
- Free access
- Local Free
- Billing unit
- Optional cloud $20/month
- Platform
- Windows / macOS / Linux
- Verified
- 17 Oct 2026
Overview
What you provide
- You provide prompts and local documents.
What you get
- You receive model responses and document answers.
Choosing this tool
Choose it when keeping document processing on a machine you control matters more than avoiding setup. The offline documentation and the cloud pricing describe different execution paths: selecting a cloud model or web search changes the privacy boundary. Verify the selected model and disable network-dependent features before treating a workflow as offline.
A decision worked through
Decision check: a downloaded local model can run without a cloud subscription; paying $20/month does not purchase faster local hardware. Start with a representative document on your own machine, then decide whether optional cloud access solves a specific gap.
LM Studio alternatives: choose an implementation route
For LM Studio, compare how models are installed and served locally; optional cloud subscriptions do not make local hardware requirements disappear.
| Product | Deployment requirements | Free allowance | Input → output | Key limit |
|---|---|---|---|---|
| LM StudioRead profileOfficial source ↗ | Desktop app with downloaded local models; choose model size against available memory. Bionic cloud access does not upgrade local hardware. | Local Free | Input: prompts and local documents. Output: model responses and document answers. Review citations and factual claims before reuse. | See the product-specific review limits below. |
| JanRead profileOfficial source ↗ | Windows/macOS/Linux desktop app with a localhost:1337 API. Download a model that fits device memory; cloud-provider mode is a separate route. | Self-managed edition | Prompts, a downloaded model or provider connection → Chat responses and local API responses | Model size must fit RAM and acceleration support. MCP tools can introduce external connections or actions; local inference alone does not make every tool offline. |
| GPT4AllRead profileOfficial source ↗ | Desktop installer and LocalDocs; CPU inference supported without a required GPU. Model files still need sufficient RAM; model licences vary. | Self-managed edition | Local models, prompts and a LocalDocs collection → Chat answers and document-retrieval responses | The application licence and individual model licences differ. Linux architecture and available memory constrain model choice; large documents still need retrieval and answer checks. |
Setup & workflow
- Download the app.
- Choose a model that fits your machine.
- Download it.
- Test a small document.
- Review the answer and active network features.
Requirements & installation
- Windows / macOS / Linux
Limits & review
- Initial model/runtime downloads require connectivity. Model size must fit your hardware. Offline capability does not establish the privacy of optional cloud services or third-party tools.
Your part
- Confirm the active model endpoint before sending a confidential document.
Sources & verification6
This profile is based on official sources, not a hands-on product test.
Vendor descriptions and demos document advertised features. Editorial guidance is based on these sources.
- Pricinglmstudio.ai
- Offline operationlmstudio.ai
- Jan official READMEgithub.com
- Jan software licencegithub.com
- GPT4All official READMEgithub.com
- GPT4All software licencegithub.com
