Keep LM Studio.
Give your local models your machines.

AI Commander is not an alternative to LM Studio, Claude Code, Codex, ChatGPT, Cursor, GitHub Copilot, Sourcegraph Cody, Grok, Gemini, Windsurf, Continue, Cline, Roo Code, Zed, OpenCode, Qwen Code, goose, Amp, Kilo Code, OpenHands, Amazon Q Developer, Augment Code, Devin, JetBrains Junie, Kiro, Warp, Factory, Crush, or Trae. It is the remote-machine harness those tools plug into over MCP: outbound-only shell, detached jobs that outlive the chat, and file transfer on hardware you already own.

Sibling harness pages: Claude Code · Codex · ChatGPT · Cursor · OpenCode · Factory · Kiro · Warp · Crush · Trae · Gemini — and the rest on Connect your AI client.

Connect LM Studio (remote MCP)

LM Studio runs models on your own hardware and, since 0.3.17, acts as an MCP host (LM Studio MCP docs) for both local and remote servers. It follows Cursor’s mcp.json notation. Open the Program tab in the right-hand sidebar, click Install → Edit mcp.json, and add AI Commander’s hosted endpoint:

{
  "mcpServers": {
    "aicommander": {
      "url": "https://aicommander.dev/mcp"
    }
  }
}

The file lives at ~/.lmstudio/mcp.json (Windows: %USERPROFILE%\.lmstudio\mcp.json). Recent LM Studio builds handle the browser-based OAuth sign-in for remote servers (MCP Integrations); if yours does not, use ?anonymous=1 for a code-only first hour or pass a Bearer API key in headers. If the file already has servers, paste only the "aicommander": { … } entry inside the existing mcpServers object.

Prefer a local stdio bridge?

{
  "mcpServers": {
    "aicommander": {
      "command": "npx",
      "args": [
        "-y",
        "@aicommander/mcp"
      ],
      "env": {
        "AICOMMANDER_SERVER": "https://aicommander.dev"
      }
    }
  }
}

LM Studio’s docs rightly warn: never install MCP servers from untrusted sources — some can run code and reach your files and network. AI Commander’s endpoint is first-party and hosted by us, and it only acts on machines where you installed the agent yourself. Keep LM Studio’s tool-call confirmations on at first, and use a model with solid tool-calling and enough context for tool output; long builds and training runs are good candidates for detached jobs so the chat does not have to wait.

Full client matrix (LM Studio, Crush, Trae, Factory, Kiro, Warp, Junie, Claude Code, Codex, ChatGPT, Cursor, Copilot, Gemini, OpenCode, goose, and more, plus REST): Connect your AI client. Menus and options can differ between LM Studio versions — check the linked LM Studio docs for your build.

What you get that the chat alone cannot

• Remote shell on Linux, macOS, and Windows — by session code or saved alias.
• Detached jobs for builds, test suites, and training on the box that has the repo.
• Fleet list + GPU facts so your model picks the right machine before it burns an hour.
• File courier (Pro) for checkpoints and diffs up to 100 MiB.
• Outbound-only agent on each machine — no inbound ports, no exposed SSH, no VPN to set up.

Not a replacement — a multiplier

FAQ