DEV Community

Cover image for Building Neo v0.1: Dual-GPU Vision Engine for Desktop Automation
Achyut Srivastava
Achyut Srivastava

Posted on

Building Neo v0.1: Dual-GPU Vision Engine for Desktop Automation

What We Built Today

Today, we shipped improvements to Neo v0.1, LuxurAI Cowork’s vision engine, which powers real-time desktop screen perception and OCR. The engine runs on our dedicated dual-GPU cluster, enabling low-latency UI element detection and task execution.

The Problem We Solved

Desktop automation requires precise screen perception, but most solutions struggle with high-DPI displays and multi-monitor setups. Neo v0.1 addresses this by combining live frame capture with advanced OCR and UI bounding box detection.

How The Architecture Works

Neo v0.1 processes desktop frames at 15 fps, detects UI elements, and uses Windows UI Automation (UIA) to execute tasks. Every high-risk action (e.g., file modifications) is gated by a human-in-the-loop safety overlay to prevent unintended changes.

Transparent Economics

LuxurAI Cowork is priced at ₹0.75/minute (3 Luxur Credits), making it accessible for developers and creators.

How We Can Make This Better

We’re actively optimizing Neo v0.1 for high-DPI screens and seeking advice from Windows/C++ developers on low-overhead frame capture. Test the beta and share your feedback: https://luxurai.in.

Try Free

Get started with 45 free Luxur Credits and explore LuxurAI Cowork: https://luxurai.in.

Top comments (0)