DeskVision App

Search images on your local PC using a local vision model with Ollama.

Features

• Native application for Windows, macOS, and Linux
• Blazing fast file search on disk (Rust-powered)
• Connect to local or remote Ollama servers via URL
• Optimized thumbnail generation for better performance
  • CPU usage: ~2% (without Ollama), Memory: 100-400MB RAM
• Automatic labeling and description generation for images
• Multi-folder support for searching, displaying, and labeling images

TODO

• File format filtering
• Search within a specific directory
• Reverse image search (search by image)
• Reverse image search within a specific directory
• Manual label editing/creation for images
• Drag & Drop folders into the app
• Thumbnail generation in a separate thread
• Improved UI, better image display, etc.
• Open larger images
• Adjustable font size
• Custom system message/prompt settings for labeling
• AI agent for enhanced image search in directories (customized prompts for the vision model)
• Ollama settings UI to pull available vision models
• Image selection, deletion, copying, and moving
• Image cropping for reverse image search
• Support for custom AI agents to automate image searching, deleting, and moving

Installation

Prerequisites: Ensure you have Ollama installed with a compatible vision model.

# Clone the repository
git clone https://github.com/yourusername/DeskVision.git
cd DeskVision

# Run the application
cargo run --release

Contribution

Feel free to submit pull requests or open issues to suggest new features or improvements!

License

MIT License