ContextBox

AI-powered context capture and organization for developers

Documentation • Installation • Quick Start

---

What is ContextBox?

ContextBox captures, organizes, and queries your working context. Take screenshots, extract web content, and ask AI questions about everything you've captured.

Key Features:

Feature	Description
Screenshot Capture	Capture screen with automatic OCR text extraction
Web Extraction	Pull content from web pages, Wikipedia, YouTube
AI Q&A	Ask questions about your context (free via GitHub Models)
Smart Search	Semantic search across all captured contexts
Cross-Platform	Works on macOS, Linux, Windows

Installation

Using pip

pip install contextbox

From source

git clone https://github.com/jayhemnani9910/contextbox.git
cd contextbox
pip install -e ".[all]"

Optional dependencies

# LLM features (GitHub Models)
pip install contextbox[llm]

# OCR support
pip install contextbox[ocr]

# YouTube extraction
pip install contextbox[youtube]

# Everything
pip install contextbox[all]

Quick Start

1. Capture your screen

contextbox capture

This takes a screenshot, extracts text via OCR, and stores it in the local database.

2. List captured contexts

contextbox list

3. Ask questions (requires GitHub token)

export GITHUB_TOKEN="your_github_token"
contextbox ask "What was I working on?"

4. Generate summaries

contextbox summarize --all-contexts

CLI Commands

Command	Description
contextbox capture	Capture screenshot and extract context
contextbox list	List all stored contexts
contextbox search <query>	Search through contexts
contextbox ask <question>	Ask questions about context (AI)
contextbox summarize	Generate context summaries (AI)
contextbox stats	Show database statistics
contextbox export	Export contexts to file

Configuration

GitHub Token (for AI features)

export GITHUB_TOKEN="ghp_xxxxxxxxxxxx"

Supported AI Models

All models are free via GitHub Models:

Model	Best For
gpt-4o-mini	Fast responses (default)
gpt-4o	Complex reasoning
meta-llama-3.1-70b-instruct	Open-source alternative

Python API

from contextbox import ContextBox

app = ContextBox()

# Capture and store
context = app.capture()
context_id = app.store_context(context)

# Search and retrieve
results = app.search("keyword")

Requirements

Platform	Screenshot Tool	OCR
macOS	Built-in	Tesseract
Linux	scrot / flameshot	Tesseract
Windows	pyautogui	Tesseract

License

MIT License

---

Documentation • Issues • Contribute