Howlite is a modern, privacy-focused Digital Asset Manager & Image Organizer for desktop. Powered by PySide6 and local AI vision (Ollama).
- 🎨 Modern Dark & Light UI: Sleek interface built with PySide6, custom titlebar, responsive grid layout, and the Syne typography.
- 🤖 Local AI Auto-Tagging & Analysis: Connects with Ollama (
llama3.2-vision,llava,moondream) to generate automatic image descriptions and smart tags locally on your GPU/CPU without sending data to external APIs. - 🐦 Twitter / X Bookmarks Scraper: Built-in scraper using Playwright to automatically sync and import your bookmarked images directly into your local library.
- 📂 Smart Organization: Custom tags with colors, smart collections, folder structures, and high-performance local search.
- 🌐 Multi-language Support (i18n): Seamless language switching (English, French, etc.).
- 📦 100% Private & Local: Powered by a local SQLite database and fast thumbnail caching.
- Python 3.10+
- (Optional) Ollama running locally for AI features.
git clone https://github.com/your-username/howlite.git
cd howlitepip install -r requirements.txtplaywright install chromiumRun the main application entry point:
python main.pyOn first launch, Howlite will prompt you to select or create a directory for your image library.
To enable automatic image tagging and AI descriptions:
- Install Ollama.
- Pull a vision model (e.g.
llama3.2-vision):ollama pull llama3.2-vision
- Make sure Ollama is running (
http://localhost:11434). - In Howlite, navigate to Settings > Ollama AI to test the connection and customize prompts.
You can package Howlite into a standalone Windows executable using PyInstaller:
python build_exe.pyThe compiled binary and resources will be generated inside the dist/ directory.
Contributions, bug reports, and feature requests are welcome!
Feel free to open an Issue or submit a Pull Request.
This project is licensed under the MIT License.