Models & providers
On-device by default. Bring Hypery, OpenAI, or a model running on your own machine.
Apple Intelligence (default)
Apple's on-device model is free, works offline, and nothing you ask leaves your Mac. It needs macOS 26 with Apple Intelligence turned on. It's small, so it's best at short answers, rewriting and extracting, and a very long conversation eventually runs out of room. When that happens, press Esc to start over.
Other providers
Open AI Models to pick a model. Built in:
| Provider | Sign-in |
|---|---|
| Hypery | Browser sign-in, or an API key |
| OpenAI | API key |
| Ollama | None, runs locally |
| LM Studio | None, runs locally |
API keys and sign-in tokens are kept in your login Keychain.
Add your own
Any OpenAI-compatible server works. Add it under ai.providers in shortcuts.json and pick it by provider:model:
{
"ai": {
"model": "work:llama-3.1-70b",
"providers": {
"work": { "url": "https://llm.example/v1", "keyEnv": "WORK_KEY" }
}
}
}Signing in
Browser sign-in uses OAuth with PKCE. Your browser passes through moo.moi/callback, which forwards it straight back to moo on your Mac. The relay keeps nothing, and the code it carries can't be used without a secret that never leaves your Mac.
If a request needs sign-in, credit or a key, moo shows the fix as the selected row, such as Sign In to Hypery or Add API Key, and asks your question again once that's done.