About ContextCaddy
ContextCaddy is an MCP server in the Search category: a clean Chrome extension for saving and searching text chunks in organized cabinets, synced in real-time with the mcp-cabinets Server.
Install
git clone https://github.com/BeckettFrey/ContextCaddyThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about ContextCaddy
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
ContextCaddy (Chrome Extension)
A sleek ui for organizing text chunks in cabinets, synced with the MCP-Cabinets Server. These chunks can then be accessed by your agent/assistant via MCP. See it in action.
[!NOTE] This repository contains only the Chrome extension code. Backend setup and services are managed separately in the MCP-Cabinets repository.
Bonsai Tree
ContextCaddy/
├── manifest.json # Extension configuration
├── background.js # Service worker & context menu
├── content.js # Web page text selection
├── sidepanel.html # Main interface
├── sidepanel.css # Compact, responsive styles
├── sidepanel.js # Core functionality with ContextCaddy class
├── icons/ # Extension icons (PNG files)
└── README.md # This file
Prerequisites
Required Backend Setup
[!IMPORTANT] ContextCaddy requires the backend services from the MCP-Cabinets repository.
-
Clone the backend repository:
git clone https://github.com/beckettfrey/mcp-cabinets.git cd mcp-cabinets -
Follow the setup instructions: Please refer to the repository's README for complete setup of both the API service and optional MCP service for AI assistant integration.
Chrome Extension
Features
Clean Interface: Modern, intuitive side panel design
Cabinet Management: Create, select, and delete cabinets
Smart Search: Search chunks within cabinets with relevance scores
Multiple Input Methods: Drag & drop or context menu
Real-time Sync: Automatic sync with MCP-Cabinets Server
Error Handling: Graceful error handling and user feedback
Configurable Settings: Adjustable text length limits
Installation
-
Ensure backend is running: MCP-Cabinets API server must be running on
localhost:8000 -
Clone this repository (it is the extension folder):
git clone https://github.com/beckettfrey/ContextCaddy.git cd ContextCaddy -
Load extension in Chrome:
- Open Chrome and navigate to
chrome://extensions/ - Enable "Developer mode" (toggle in top right)
- Click "Load unpacked" and select your
ContextCaddyfolder - The ContextCaddy icon will appear in your toolbar
- Open Chrome and navigate to
Usage
Opening the Side Panel
- Click the ContextCaddy extension icon, OR
- Right-click the icon and select "Open side panel"
Managing Cabinets
- Create: Click "+ New" and enter a cabinet name (1-100 characters)
- Select: Click on any cabinet in the list to activate it
- Delete: Select a cabinet, then click "Delete" (requires confirmation)
Adding Text Chunks
Method 1: Drag & Drop
- Select a cabinet
- Highlight text on any webpage
- Drag the selected text to the drop zone in the side panel
Method 2: Context Menu
- Select a cabinet (optional - will prompt if none selected)
- Right-click on highlighted text on any webpage
- Choose "Add to ContextCaddy Cabinet"
Searching Chunks
- Select a cabinet
- Enter search terms in the search box
- Press Enter or click "Go"
- Results show with relevance scores and source URLs
- Adjust "Results" count and "Min Score" as needed
Settings
- Click the ⚙️ settings button to configure text length limits
- Min Text Length: Minimum characters required for text chunks (default: 10)
- Max Text Length: Maximum characters allowed for text chunks (default: 5000)
Development
To modify the extension:
- Make changes to the relevant files
- Go to
chrome://extensions/ - Click the refresh icon on the ContextCaddy extension
- Test your changes
Additional Notes
- Backend: MCP-Cabinets API service
- Browser: Chrome with Extension Manifest V3 support
- No build process: Direct file loading for development
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.