Data in OWA

Desktop AI needs high-quality, synchronized multimodal data: screen captures, mouse/keyboard events, and window context. OWA provides the complete pipeline from recording to training.

Quick Start

# 1. Record desktop interaction
$ ocap my-session.mcap

# 2. Process to training format
$ python scripts/01_raw_to_event.py --train-dir ./

# 3. Train your model
$ python train.py --dataset ./event-dataset

📖 Detailed Guide: Complete Quick Start Tutorial

Documentation

🚀 Getting Started

• Why OWAMcap? - Understand the problem and solution
• Recording Data - Capture desktop interactions with ocap
• Exploring Data - View and analyze your recordings

📚 Technical Reference

• OWAMcap Format Guide - Complete technical specification
• Data Pipeline - Transform recordings to training-ready datasets

🛠️ Tools

• Data Viewer - Web-based visualization tool
• Data Conversions - Convert existing datasets (VPT, CS:GO) to OWAMcap

🤗 Community Datasets

Browse Datasets: 🤗 HuggingFace

• Standardized Format: All datasets use OWAMcap for seamless integration
• Interactive Preview: Hugging Face Spaces Visualizer