Skip to content

Repository files navigation

AI Shortcuts

1276@3x

This app provides Siri Shortcuts to interact with AI.

The app itself provides configuration for the shortcuts; the actual shortcut building blocks are available in the Shortcuts app on macOS.

This app is compatible with macOS.

Under the hood, the app uses MacPaw/OpenAI.

-> AppStore

Privacy

See PRIVACY_POLICY.md.

Shortcuts (App Intents)

These are the actions the app exposes to the Shortcuts app.

1. Ask AI

Intent: Ask AI

  • Description: Sends a prompt (optionally with images) and returns the text response.
  • Input:
    • Prompt (String)
    • System Prompt (Optional String)
    • Model (Optional String; defaults to the app’s chat default)
    • Images (Optional [File], images only, max 20)
  • Output: String

2. Generate Image

Intent: Generate Image

  • Description: Creates an image based on a prompt.
  • Input:
    • Prompt (String)
    • Model (Optional String; defaults to the app’s image default)
    • Size (Enum: Auto, 1024×1024, 1024×1536, 1536×1024)
  • Output: File (PNG)
  • Note: Requires the official OpenAI endpoint (or a provider that supports POST /v1/images/generations).

3. Audio

Intent: Transcribe Audio

  • Description: Converts an audio file into text using AI transcription.
  • Input:
    • Audio File (File, audio)
    • Model (Optional String; defaults to the app’s transcription default)
    • Language (Optional String; ISO-639-1 like en, empty = auto-detect)
  • Output: String (Transcription)
  • Supported file types: mp3, mp4, m4a, wav, webm, mpeg/mpga, ogg/oga, flac (defaults to m4a if unknown).
  • Note: Requires the official OpenAI endpoint (or a provider that supports POST /v1/audio/transcriptions).

Intent: Read Text Aloud (TTS)

  • Description: Generates spoken audio from text input.
  • Input:
    • Text (String)
    • Model (Optional String; defaults to the app’s TTS model, falling back to tts-1)
    • Voice (Enum: Alloy, Echo, Fable, Onyx, Nova, Shimmer)
    • Speed (Double; clamped to 0.25–4.0, default 1.0)
  • Output: File (MP3)
  • Note: Requires the official OpenAI endpoint (or a provider that supports POST /v1/audio/speech).

Config

  • API Key: Stored securely (Keychain + iCloud Keychain sync).
  • Endpoint Settings: Host, Base Path, Scheme, Port (leave empty to use https://api.openai.com/v1).
  • Default Models: Optional per-feature defaults used when a shortcut action’s Model parameter is left empty.

Local build scripts

This repo includes a small set of scripts to help avoid committing local signing settings.

  • One-time (per clone): ./scripts/install-githooks.sh (enables the pre-commit hook to run ./clean.sh before commits)
  • Create local env: cp .env.example .env
  • Configure signing: ./configure.sh
  • Build + run (macOS): ./build-macos.sh

^^

About

AI Actions for Apple Shortcuts

Resources

Stars

1 star

Watchers

0 watching

Forks

Contributors

Languages