Offline Mac AI Video Creator

I need a native macOS application that runs entirely on my 2025 15-inch MacBook Air (M4, 16 GB unified memory) with zero cloud calls or subscriptions. Everything—Text-to-Image, Image-to-Video, and Video-to-Video generation—must execute locally, taking full advantage of the M-series Neural Engine, Metal, Core ML, or any other on-device acceleration you prefer.

Interface
Please keep the UI minimal and straightforward so I can focus on creating, not navigating menus. A single window with clear tabs or panels for each mode would be ideal.

Core workflow
• Text-to-Image: let me switch among realistic, cartoon, or abstract styles at runtime.
• Image-to-Video & Video-to-Video: generate clips up to 10 seconds at 24–30 fps.
• Built-in timeline tools for trimming, extending, and laying a background-music track (simple volume and fade controls are enough).

Deliverables
1. Signed, notarised .app bundle ready to drag-and-drop into /Applications.
2. All model files stored within the app package or its local support folder—no external downloads after install.
3. Source code and a brief build/read-me so I can recompile if Apple ships a new SDK.
4. Quick-start guide that shows: installation, generating the first clip, exporting . mp4 or ProRes.

Acceptance criteria
• Launches and processes media in Airplane Mode.
• Generates a 10-second 1080p clip from a text prompt in under five minutes on my hardware.
• Memory footprint stays within 16 GB (swap bursts are fine, but no crashes).
• UI matches the minimal style described above.

When you reply, please give me a single price quote and an estimated timeline from kickoff to notarised build. Feel free to outline which on-device models or libraries you plan to integrate (e.g., Stable Diffusion, LLaVA, MLV-AE, SwiftUI, Metal Performance Shaders).