agentAR: Creating Augmented Reality Applications with Tool-Augmented LLM-based Autonomous Agents
summary
agentAR uses a tool-augmented LLM autonomous agent plus a reusable tool library (object detection, speech recognition, pose tracking, rendering, UI modules) to plan and execute end-to-end AR apps from dialogue, producing a semi-structured plan and wiring front-end and back-end modules automatically. Six diverse apps appear in the case study and a 12-participant user study reports weeks to minutes speedups (one prior app rebuilt in under 10 minutes) and significantly reduced user effort; the system is modular, runtime-editable, and the team plans to open-source the workflow.
AI-written summary · about
related talks
33mAR as a Feature: How to Supercharge Products Using Augmented Reality
12mContextual Computing & AR to Build the 21st Century Workforce
15mGhostAR: A Time-space Editor for Embodied Authoring of Human-Robot Collaborative Task with ...
34mAI Agents: Transforming Autonomous Decision-Making
17mAR in Enterprise Solutions? This is What You Need to Know.
49mAI Agents: Collaboration in Real-Life Applications
12mAR for the Enterprise at AWE 2015
16mWhy 2017 is the Year AR Will Move From Hollywood Fantasy