KASCA-esque

A dual-endpoint web application for synchronized keystroke and audio recording. The system consists of a keyboard input endpoint and a control/monitoring endpoint that can record audio while capturing keystrokes from remote keyboard clients.

Documentation

• SPEC.md - Full specification and architecture
• PROTOCOL.md - WebSocket protocol specification

Installation

npm install

Development

Start the development server:

npm run dev

Note: In development mode, the WebSocket server is not available. You must build and run in production mode to use WebSocket features.

Production

Build the application:

npm run build

Run the production server (with WebSocket support):

npm start
# or
npm run preview

The server will start on http://localhost:3000 by default.

Usage

1. Home Page (/): Navigation to all endpoints
2. Keyboard Endpoint (/keyboard): Open this on devices where you want to capture keystrokes
3. Control Panel (/control): Monitor connected keyboards, select one, and record audio synchronized with keystrokes
4. Playback Panel (/control/playback): View, play, and delete recordings

WebSocket Endpoints

• ws://localhost:3000/ws/keyboard - Keyboard clients connect here
• ws://localhost:3000/ws/control - Control clients connect here

Recordings

Recordings are stored in ./recordings/ directory with:

• recording_YYYYMMDD_HHMMSS.json - Metadata and keystroke data
• recording_YYYYMMDD_HHMMSS.webm - Audio file

Architecture

Built with:

• Frontend: Svelte 5 + SvelteKit
• Backend: Node.js with Express
• WebSockets: ws library
• Audio: MediaRecorder API (WebM format)
• Styling: Dark theme with green highlights (terminal/hacker aesthetic)