Data in OWA: Complete Desktop Agent Data Pipeline

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: Record → Train in 3 Steps

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

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

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

📖 Detailed Guide: Complete Quick Start Tutorial - Step-by-step walkthrough with examples and troubleshooting

The OWA Data Ecosystem

🎯 Getting Started

New to OWA data? Start here:

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

📚 Technical Reference

Deep dive into the format and pipeline:

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

🛠️ Tools & Ecosystem

• Data Viewer - Web-based visualization tool
• Comparison with LeRobot - Technical comparison with alternatives

🤗 Community Datasets

Browse Available Datasets: 🤗 datasets?other=OWA

• Growing Collection: Hundreds of community-contributed datasets
• Standardized Format: All use OWAMcap for seamless integration
• Interactive Preview: Hugging Face Spaces Visualizer
• Easy Sharing: Upload recordings directly with one command

🚀 Impact: OWA has democratized desktop agent data, growing from zero to hundreds of public datasets in the unified OWAMcap format.