This program is an AI-powered screen activity logger that captures screenshots at regular intervals, analyzes them using OpenAI/Claude/Your Model, and logs the descriptions. It also allows users to ask questions about their activity log, providing a unique way to track and understand your computer usage.
- Automatic screenshot capture at customizable intervals, interpreted by vision model of your choice.
- Q&A mode to ask questions about your activity.
- Privacy model support (via LM Studio or Ollama) to gatekeep what gets sent to stronger models, or use all local models.
- Clone this repository
- Install the required packages:
pip install -r requirements.txt
- Create a
.env
file in the project root and add your OpenAI API key:OPENAI_API_KEY=your_api_key_here
- Use
config.json
to configure your models and other settings.
-
Run the program:
python main.py
-
The program will start capturing screenshots and logging activity.
-
Press Enter at any time to pause capturing and enter inquiry mode.
-
In inquiry mode, you can:
- Ask questions about your activity
- Type 'exit' to quit the program
- Type 'reset' to start a new session
- Type 'continue' to resume screen capturing
This program uses your screen content and sends it to OpenAI for analysis. Make sure you're comfortable with this and comply with all relevant privacy policies and regulations.