UFO^3: Weaving the Digital Agent Galaxy
Chaoyun Zhang, Liqun Li, He Huang, Chiming Ni, Bo Qiao, Si Qin, Yu Kang, Minghua Ma, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
2025-11-18
Summary
This paper introduces UFO^3, a new system designed to let AI agents, powered by large language models, work seamlessly across all your devices – phones, computers, servers, and even smaller edge devices – as if they were one unified system.
What's the problem?
Currently, AI agents are usually limited to working within a single device or operating system. This makes it difficult to create complex tasks that require using multiple devices together, and often requires a lot of manual setup and intervention from the user. It's like trying to build with LEGOs where the bricks from different sets don't connect well.
What's the solution?
UFO^3 solves this by treating each user request as a 'TaskConstellation,' which is essentially a map of smaller, individual tasks ('TaskStars') and how they depend on each other ('TaskStarLines'). This map can spread across all connected devices and automatically adjusts as tasks are completed. A 'Constellation Orchestrator' manages everything, ensuring tasks run safely and efficiently, and a special communication system ('Agent Interaction Protocol') keeps everything connected with minimal delay. It's like a project manager that dynamically assigns tasks to the best available team members (devices).
Why it matters?
This work is important because it moves us closer to a future where AI agents can truly assist us in a smart and integrated way, using all our devices to their full potential. By breaking down the barriers between devices, UFO^3 allows for more complex, efficient, and reliable AI-powered workflows, making our technology more helpful and adaptable to our needs.
Abstract
Large language model (LLM)-powered agents are transforming digital devices from passive tools into proactive intelligent collaborators. However, most existing frameworks remain confined to a single OS or device, making cross-device workflows brittle and largely manual. We present UFO^3, a system that unifies heterogeneous endpoints, desktops, servers, mobile devices, and edge, into a single orchestration fabric. UFO^3 models each user request as a mutable TaskConstellation: a distributed DAG of atomic subtasks (TaskStars) with explicit control and data dependencies (TaskStarLines). The TaskConstellation continuously evolves as results stream in from distributed devices, enabling asynchronous execution, adaptive recovery, and dynamic optimization. A Constellation Orchestrator} executes tasks safely and asynchronously while applying dynamic DAG updates, and the Agent Interaction Protocol (AIP) provides persistent, low-latency channels for reliable task dispatch and result streaming. These designs dissolve the traditional boundaries between devices and platforms, allowing agents to collaborate seamlessly and amplify their collective intelligence. We evaluate UFO^3 on NebulaBench, a benchmark of 55 cross-device tasks across 5 machines and 10 categories. UFO^3 achieves 83.3% subtask completion, 70.9% task success, exposes parallelism with an average width of 1.72, and reduces end-to-end latency by 31% relative to a sequential baseline. Fault-injection experiments demonstrate graceful degradation and recovery under transient and permanent agent failures. These results show that UFO^3 achieves accurate, efficient, and resilient task orchestration across heterogeneous devices, uniting isolated agents into a coherent, adaptive computing fabric that extends across the landscape of ubiquitous computing.