A minimalist client solution for Ollama written in python with ability to capture screen to use with vision models.
- Screenshot capturing with multi-monitor support
- Markdown / Code blocks support
- Easy tagging for models that are vision enabled
- 3 display modes (Collapsed, Minimal and Expanded)
- @ on start of prompt to easily choose the model
- Drag and Drop / Paste images to prompt
To-Do:
- Context persistency between sessions
- Refactor and removal of unecessary or redundant code
git clone https://github.com/fredconex/PixelLlama.git
cd ./PixelLlama
./run.bat
Execution of run.bat will create a new environment, install the required packages automatically and launch PixelLlama.
