Feature
AI chat, Quick Fix and agents
Last reviewed 17 October 2026
SuperCmd puts AI one keystroke away without selling you a subscription. You connect the provider you already pay for, sign in with a ChatGPT account, or run models on the Mac itself with Apple Intelligence, local MLX, Ollama or LM Studio. The same models power Quick AI on selected text, Quick Fix rewrites, a dedicated chat window, agents that plan and use tools, and a live voice mode. Nothing is routed through SuperCmd's servers; requests go directly from your Mac to the provider you chose.
Supported providers
Every surface can use a different provider and model, so a cheap fast model can handle Quick Fix while a frontier model runs agents. Each AI feature has its own switch and can be turned off entirely.
- OpenAI
- ChatGPT account sign-in (no separate API key)
- Anthropic
- Google Gemini
- OpenRouter
- Mistral
- xAI Grok
- ElevenLabs
- Apple Intelligence (on-device)
- Local MLX models
- Ollama
- LM Studio
- Any OpenAI-compatible endpoint
AI surfaces
| Surface | What it does |
|---|---|
| Quick AI | Ask a question from the launcher and get an answer inline, with the selected text or clipboard as context. |
| Quick Fix | Select text in any app, press the hotkey, and have it rewritten, fixed or translated in place. |
| AI Chat window | A separate window for long conversations that stays open while you work. |
| Agents | A harness that plans, calls tools such as files, shell and web, and reports back when the job is done. |
| Global AI prompt | A system-wide prompt bound to a hotkey, with the frontmost app's selection as input. |
| GPT Realtime | A live voice conversation with OpenAI's realtime models. |
Privacy
SuperCmd stores API keys in the macOS Keychain and sends prompts straight to the configured provider. On-device options, Apple Intelligence and local MLX, never leave the Mac. SuperCmd does not proxy, log or train on your prompts.
Frequently asked questions
Do I need an AI subscription to use SuperCmd's AI features?
No. SuperCmd has no AI subscription. Use your own API key, sign in with an existing ChatGPT account, or run models on-device with Apple Intelligence, local MLX, Ollama or LM Studio.
Can SuperCmd use local models?
Yes. SuperCmd supports Apple Intelligence on-device, local MLX models, Ollama and LM Studio, so AI features can run without an internet connection.
Which AI providers does SuperCmd support?
OpenAI, ChatGPT account sign-in (no separate API key), Anthropic, Google Gemini, OpenRouter, Mistral, xAI Grok, ElevenLabs, Apple Intelligence (on-device), Local MLX models, Ollama, LM Studio, Any OpenAI-compatible endpoint.
Is SuperCmd's AI the same as Raycast AI?
Raycast AI is a hosted service included with Raycast Pro. SuperCmd connects to providers you choose and includes the same kinds of surfaces: inline answers, text rewriting, a chat window and agents.