Keep ChatGPT.
Give it your machines.

AI Commander is not an alternative to ChatGPT, OpenAI, GPT, Codex, Claude Code, Cursor, or Grok. It is the remote-machine harness those tools plug into over MCP — including OpenAI submodels such as Astra (gpt-6-astra): outbound-only shell, detached jobs that outlive the chat, and file transfer on hardware you already own.

Positioning in one line: ChatGPT / GPT / Astra stay 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: Claude Code · Cursor · Grok · Codex.

Connect ChatGPT / OpenAI (MCP connector)

ChatGPT (and OpenAI GPT apps that support custom connectors) talk to AI Commander as a custom MCP connector — a remote HTTPS endpoint. Works the same whether you are on ChatGPT, GPT, or an OpenAI submodel such as Astra (gpt-6-astra).

In ChatGPT, enable Developer mode in Settings → Security and login. Then open ChatGPT Plugins, select plus (+), enter a name and description, and add the public MCP URL under Connection:

URL: https://aicommander.dev/mcp
Auth: leave empty — the connector auto-registers and you complete a one-time browser sign-in.

ChatGPT custom connector

https://aicommander.dev/mcp

For a code-only first hour without sign-in, use https://aicommander.dev/mcp?anonymous=1. No connector support? Drive a machine over plain REST POST /api/v1/exec — see Chat apps and REST API.

Full client matrix (ChatGPT, Codex, Claude Code, Cursor, Grok, REST): Connect your AI client.

OpenAI models, GPT, and Astra

AI Commander does not care which OpenAI model is answering — ChatGPT, GPT, or a submodel keyword like Astra / gpt-6-astra (the same kind of label as Opus or Fable). Your model stays the brain; this page is the harness that gives it your machines.

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, training, and headless agents on the box that has the repo.
Fleet list + GPU facts so ChatGPT picks the right machine before it burns an hour.
File courier (Pro) for checkpoints and diffs up to 100 MiB.

Not a replacement — a multiplier

You already use	AI Commander adds
ChatGPT / OpenAI / GPT / Astra	Custom MCP connector or plain REST `/api/v1/exec`
Codex CLI	`codex mcp add aicommander --url …` → remote exec + jobs
Grok / `grok` CLI	`grok mcp add --transport http …` → remote exec + jobs
Claude Code	MCP remote exec + jobs on your fleet
Cursor	`npx @aicommander/mcp` stdio bridge → remote exec + jobs

FAQ

Is AI Commander an alternative to ChatGPT or OpenAI?

No. AI Commander is a remote-machine harness. You keep ChatGPT, GPT, Codex, Claude Code, Cursor, or Grok as the AI — including OpenAI submodels such as Astra (gpt-6-astra). 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 ChatGPT to AI Commander?

In ChatGPT, enable Developer mode in Settings → Security and login. Then open ChatGPT Plugins, select plus (+), enter a name and description, and add the public MCP URL under Connection. URL: https://aicommander.dev/mcp — leave Auth empty for the one-time browser sign-in. Use https://aicommander.dev/mcp?anonymous=1 for a code-only first hour. No connector? Drive machines over REST POST /api/v1/exec. See /howto/#chat.

Does this work with GPT and Astra (gpt-6-astra)?

Yes. AI Commander is model-agnostic. ChatGPT, GPT, and OpenAI submodels such as Astra (gpt-6-astra) stay the brain; the harness adds hands on your fleet over MCP or REST — same connector URL either way.

Next steps

Install the agent on a machine you control, add the ChatGPT custom connector, then ask for a real task on a remote box.

Install the agent Setup guides Remote coding agents

Keep ChatGPT.Give it your machines.