A minimalist client solution for Ollama / OpenAI API written in python with ability to capture screen to use with vision models.
- Screenshot capturing with multi-monitor support
- Drag and Drop / Paste images to prompt
- 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
- Chat history persistence
- System prompt browser
- Theme support
- Compatible with Ollama / OpenAI API
To-Do:
- Refactor and removal of unecessary or redundant code (30-40% done)
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.
In the terminal
git clone https://github.com/fredconex/PixelLlama.git # clone repo
cd PixelLlama
python3 -m venv .venv #create virtual env
source .venv/bin/activate # activate virtual env
pip install -r requirements.txt #install dependencies
python main.py #start script
