Timmy Time

Sovereign AI command center for iPad. Chat with Timmy, share images, voice, links, and documents. All processing happens on your Mac — the iPad is the eyes, ears, and hands; the Mac is the brain.

Architecture

iPad (native SwiftUI app)          Mac (headless server)
┌─────────────────────────┐        ┌──────────────────────────┐
│  Chat UI                │        │  FastAPI backend          │
│  Camera / Photos        │──────→ │  POST /api/v1/chat       │
│  Microphone             │  Tail- │  POST /api/v1/upload     │
│  File picker            │  scale │  WS   /api/v1/chat/ws    │
│  URL sharing            │        │                          │
│                         │ ←───── │  Media pipeline:         │
│  Streaming responses    │        │    Images → Vision model │
│  Voice playback         │        │    Audio  → Whisper STT  │
│  Rich media display     │        │    URLs   → Web extract  │
│                         │        │    Docs   → OCR/parse    │
│                         │        │                          │
│                         │        │  Timmy agent (Agno)      │
│                         │        │  Ollama (qwen3:30b)      │
└─────────────────────────┘        └──────────────────────────┘

Requirements

iPad Pro with iPadOS 26.1+
Mac with macOS 14+ running the Timmy dashboard server
Tailscale connecting both devices
Xcode 26+ on Mac for building

MVP Features

Project structure
Chat screen with streaming responses
Image attachment (camera + photo library)
Voice input (mic → Whisper on Mac)
URL link sharing and summarization
Persistent chat history
Server-side API endpoints for the app

Future

Apple Pencil drawing/annotation
On-device models (Core ML)
LiDAR / ARKit spatial awareness
Siri Shortcuts integration
Push notifications
Video processing