Installation

Installation & First Launch

This guide details system requirements, supported platforms, installation procedures, and the first-run onboarding setup for Berry AI Studio.

---

1. System Requirements

Berry AI Studio uses an ultra-efficient native architecture powered by Tauri v2, Rust, and SQLite WAL. It runs smoothly on modest hardware while fully utilizing multi-core workstations and NVMe storage for large libraries (50,000 to 500,000+ files).

Minimum Hardware Requirements

• CPU: Dual-core x86_64 or ARM64 processor (Intel Core i3 / AMD Ryzen 3 / Apple M1 or newer).
• RAM: 4 GB RAM (8 GB+ recommended for running local CLIP/WD14 ONNX models).
• Storage: ~150 MB for application installation; additional space for thumbnails (configurable 2 GB default LRU cache) and media files.
• Display Resolution: 1280 × 800 minimum viewport (responsive down to 960 × 640).

Supported Operating Systems

Operating System	Supported Versions	Architecture	Package Types
Windows	Windows 10 (1809+) & Windows 11	x86_64 (64-bit)	Standard Installer (.exe), Portable (.zip)
macOS	macOS 12 (Monterey) or later	aarch64 (Apple Silicon M1/M2/M3/M4) & x86_64 (Intel)	Disk Image (.dmg), Universal Binary
Linux	Ubuntu 20.04+, Debian 11+, Fedora 36+, Arch Linux	x86_64	AppImage (.AppImage), Debian package (.deb)

---

2. Installation Procedures

Download official production packages from the GitHub Releases Page or the Official Website.

Windows

1. Standard Installer (Berry-AI-Studio_Windows_x64.exe):
   • Double-click the installer executable.
   • Follow the setup wizard to choose the installation location and create desktop/start menu shortcuts.
   • The installer automatically manages desktop shortcuts and register protocol handlers.
2. Portable ZIP (Berry-AI-Studio_Windows_x64.zip):
   • Extract the .zip archive to your preferred drive (e.g., an external NVMe SSD or portable drive).
   • Run berry-ai-studio.exe directly without administrative privileges.

macOS

1. Download the disk image corresponding to your CPU:
   • Apple Silicon (M1/M2/M3/M4): Berry-AI-Studio_macOS_aarch64.dmg
   • Intel Core: Berry-AI-Studio_macOS_x64.dmg
2. Open the .dmg file and drag Berry AI Studio into your /Applications folder.
3. Packages are signed and notarized by Apple Gatekeeper. On first launch, launch from Applications or Spotlight.

Linux

1. AppImage (Berry-AI-Studio_Linux_x64.AppImage):
   • Make the binary executable:

     chmod +x Berry-AI-Studio_Linux_x64.AppImage
     ./Berry-AI-Studio_Linux_x64.AppImage

2. Debian / Ubuntu (Berry-AI-Studio_Linux_x64.deb):
   • Install via dpkg or apt:

     sudo dpkg -i Berry-AI-Studio_Linux_x64.deb
     sudo apt-get install -f # Resolve any missing webkit2gtk dependencies

---

3. First-Run Onboarding Wizard

When you launch Berry AI Studio for the first time, the interactive Onboarding Wizard (OnboardingModal.vue) appears automatically to guide you through setup.

flowchart TD
    A[Launch Berry AI Studio] --> B[Welcome & Concept Introduction]
    B --> C[Local AI Tool Detection]
    C --> D{Found Local Installations?}
    D -- Yes --> E[One-Click Add SD WebUI / ComfyUI / Fooocus]
    D -- No --> F[Manual Folder Selection]
    E --> G[Configure Ingestion Mode: Link vs Pipeline]
    F --> G
    G --> H[Initial Fast Indexing]
    H --> I[Ready to Explore Studio]

Loading

Onboarding Steps:

1. Welcome Screen: Introduces the 3 core pillars:
   • High-speed local indexing with lossless generation metadata extraction.
   • Intelligent poker-card burst stacking and side-by-side comparison.
   • 100% offline privacy with zero telemetry.
2. Local AI Engine Auto-Detection:
   • Berry probes common local directories across all drives (e.g., C:\, D:\, /home/) looking for outputs from:
     • AUTOMATIC1111 / SD.Next (outputs/txt2img-images, outputs/img2img-images)
     • ComfyUI (ComfyUI/output)
     • Fooocus (Fooocus/outputs)
     • InvokeAI (invokeai/outputs)
   • If detected, you can connect them with a single click as AIGC Pipelines or External Links.
3. Select Storage Mode:
   • Choose how Berry interacts with your files (read more in Folder Modes & Importing).
4. Completion:
   • Berry initializes the local SQLite database (berry.db) in Write-Ahead Logging (WAL) mode, starts background folder scanning, and brings you directly into the main studio gallery.

---

4. Application Storage & Data Directory

Berry AI Studio stores all library indexes, caches, and configuration locally in your user profile:

• Windows: %APPDATA%\com.berryuiki.berryaistudio\ (e.g., C:\Users\<User>\AppData\Roaming\com.berryuiki.berryaistudio\)
• macOS: ~/Library/Application Support/com.berryuiki.berryaistudio/
• Linux: ~/.config/com.berryuiki.berryaistudio/

Directory Contents:

• berry.db: The primary SQLite database containing all metadata, ratings, tags, albums, and stack relationships.
• berry.db-wal & berry.db-shm: SQLite WAL journal files.
• config.json: Application settings (theme, view mode, thumbnail resolution, interop URLs).
• thumbnails/: High-efficiency WebP thumbnail cache organized by {file_id}_{mtime}_{edge}.webp.
• models/: Local ONNX AI weights for CLIP, SigLIP, and WD14 Danbooru auto-taggers.

Tip

You can instantly open these locations at any time from Settings > Storage & About using the dedicated "Open Folder" buttons.