Keep goose.
Give it your machines.

AI Commander is not an alternative to goose, Claude Code, Codex, ChatGPT, Cursor, Grok, Gemini, Windsurf, Continue, Cline, Roo Code, Zed, OpenCode, or Qwen Code. 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.

Quick start:
  1. Install the host agent on the machine you want your AI to control. It shows a private machine code.
  2. Connect your AI: paste this prompt, then replace AIC-… with your code — Help me connect this AI client to AI Commander using https://aicommander.dev/start.md. Then check the status of my machine: AIC-… — or add the connector URL https://aicommander.dev/mcp and sign in with an email code (per-client steps).
Positioning in one line: goose stays the brain; AI Commander is the hands on the box — servers, GPU rigs, NAS, laptops — with no inbound ports and no exposed SSH.

Sibling harness pages: Cursor · Aider · Cline · Roo Code · Zed · OpenCode · Qwen Code · goose · Continue · Windsurf · Claude Code · Grok · Codex · ChatGPT · Gemini.

Connect goose (MCP Streamable HTTP)

goose (docs) is Block’s open-source AI agent with native remote MCP extensions over Streamable HTTP (SSE transport was removed — use Streamable HTTP). Point it at AI Commander’s hosted MCP endpoint:

goose CLI — one session
$ goose session --with-streamable-http-extension "https://aicommander.dev/mcp"

Persist the extension with interactive configure, or edit ~/.config/goose/config.yaml (goose uses uri and type: streamable_http — not url):

goose configure
$ goose configure # Add Extension → Remote Extension (Streamable HTTP) # Name: aicommander # URI: https://aicommander.dev/mcp
~/.config/goose/config.yaml
extensions: aicommander: enabled: true type: streamable_http name: aicommander description: Remote machines via AI Commander uri: "https://aicommander.dev/mcp" timeout: 300

Full client matrix (goose, Claude Code, Codex, ChatGPT, Cursor, Grok, Gemini, OpenCode, Qwen Code, Continue, Windsurf, Cline, Roo Code, Zed, REST): Connect your AI client. Extension docs: Using extensions · CLI commands.

First connect may open a one-time browser sign-in (OAuth Dynamic Client Registration). For a code-only first hour without that flow, append ?anonymous=1 to the MCP URI. For optional account features (saved aliases), add a header in config: headers: { Authorization: "Bearer <your-key>" } (or enter the same header when prompted by goose configure).

What you get that the chat alone cannot

Not a replacement — a multiplier

You already useAI Commander adds
goosegoose session --with-streamable-http-extension "https://aicommander.dev/mcp"
Qwen Codeqwen mcp add --transport http aicommander https://aicommander.dev/mcp
OpenCodeopencode mcp add aicommander --url https://aicommander.dev/mcp
Claude Codeclaude mcp add --transport http … → remote exec + jobs
Codex / ChatGPTSame MCP endpoint or REST /api/v1/exec
Cursornpx @aicommander/mcp stdio bridge → remote exec + jobs
Grok / Geminigrok mcp add --transport http … / gemini mcp add
Cline / Roo Code / ZedMCP settings / context_servers + bridge or remote URL
Continue / Windsurfstdio bridge in Continue / Windsurf MCP config

How goose handles MCP servers

Checked against goose docs: Using extensions in November 2026. Client behaviour changes between releases; the official docs win if they disagree.

A workflow that suits goose

goose works autonomously by default. That suits jobs you start and walk away from, as long as you set the permission mode on purpose.

  1. Add the aicommander extension with a generous timeout, and pick a permission mode that asks before tools run, at least until you trust the flow.
  2. Write a short prompt you can reuse: "On the NAS, check disk usage and SMART status, then start a scrub as a background job if anything looks wrong."
  3. goose calls remote_exec for the quick checks and remote_job_start for the scrub. You come back later and ask for remote_job_logs.

Limits and gotchas in goose

FAQ

Is AI Commander an alternative to goose?
No. AI Commander is a remote-machine harness. You keep goose (or Claude Code, Codex, ChatGPT, Cursor, Grok, Gemini, Windsurf, Continue, Cline, Roo Code, Zed, OpenCode, Qwen Code) as the AI. AI Commander gives those tools secure remote shell, detached jobs, and file transfer on machines you own — over MCP or REST.
How do I connect goose to AI Commander?
Add AI Commander as a Remote Extension (Streamable HTTP). Quick session: goose session --with-streamable-http-extension "https://aicommander.dev/mcp". Or run goose configure → Add Extension → Remote Extension (Streamable HTTP) and enter that URI. Or add type: streamable_http with uri: "https://aicommander.dev/mcp" under extensions in ~/.config/goose/config.yaml. First connect may open a one-time browser sign-in (OAuth Dynamic Client Registration). Use ?anonymous=1 for a code-only first hour. Optional: set headers Authorization: Bearer <your-key> for account aliases. See /howto/ for the full matrix.
How is this different from goose running locally only?
Local goose is the brain in your terminal or Desktop. AI Commander adds hands on servers, GPU rigs, NAS, and other laptops you own — outbound-only, no inbound ports, no exposed SSH — so goose can run shell, detached jobs, and file transfer on those machines over MCP.
Will goose run commands on my machines without asking?
It can. goose is autonomous by default. Choose a permission mode that requires approval if you want to confirm each remote_exec.
Why is AI Commander missing from my new goose session?
It was probably enabled only for one session, with /extension or the --with-streamable-http-extension flag. Add it as a regular extension in config.yaml.

Next steps

Install the agent on a machine you control, connect goose via Streamable HTTP to https://aicommander.dev/mcp, then ask for a real task on a remote box.