SwiftVision.ai

Image Annotation Studio — A fast, keyboard-driven bounding box annotation tool for creating YOLO-format datasets. 100% offline, no cloud, no account required.

✨ Features

• 🖱️ Draw boxes with mouse — drag corners to resize
• ⌨️ Arrow keys to nudge boxes (Shift = 10px steps)
• 🔢 Number keys 1-9 to apply classes instantly
• ↩️ Undo / Redo with Ctrl+Z / Ctrl+Y
• 💾 Auto-save YOLO .txt files — only for annotated images
• 📋 Copy labels from previous image in one click
• 🎨 SwiftVision green branding with custom logo + favicon
• 🖼️ Zoom, pan, hover crosshair guides for precision
• 🗑️ Right-click a box to delete instantly
• 🎯 Full-screen canvas with sidebars

📦 Installation

pip install swiftvision

Requirements: Python 3.8+, Pillow

Linux users: If Tkinter is missing, install it first:

sudo apt install python3-tk      # Debian/Ubuntu
sudo dnf install python3-tkinter # Fedora

🚀 Quick Start

swiftvision

Then:

1. Click Load Folder and pick your images directory
2. Add class names in the right sidebar (or type + Enter)
3. Draw boxes by dragging on the image
4. Assign a class via single-click or number keys 1-9
5. Use Ctrl+Z / Ctrl+Y for undo/redo
6. Click Save All Labels to write .txt files

⌨️ Keyboard Shortcuts

🖱️ Mouse Controls

📂 Export Formats

• YOLO — .txt files with normalized class x y w h
• JSON — full project export with metadata
• Project file — .json for resuming work

Empty images (no labels) do not produce .txt files.

🛠️ Development

git clone https://github.com/yourusername/swiftvision.git
cd swiftvision
pip install -e .
swiftvision

📜 License

MIT — see LICENSE file.

🙌 Contributing

Issues and pull requests welcome at github.com/hassaanswift/swiftvision.

Built with ❤️ by SwiftVision AI — Vision | Innovate | Engineer