Agent JIT Compilation for Latency-Optimizing Web Agent Planning and Scheduling

TL;DR

This paper introduces agent JIT compilation, a method to reduce latency and errors in web automation agents by compiling task descriptions into executable code with integrated planning and scheduling.

Contribution

It presents a novel JIT compilation framework with planning, scheduling, and invariant protocols to improve web agent performance and accuracy.

Findings

10.4× speedup in web application tasks

28% accuracy improvement over Browser-Use

2.4× speedup with 9% accuracy gain over OpenAI CUA

Abstract

Computer-use agents (CUA) automate tasks specified with natural language such as "order the cheapest item from Taco Bell" by generating sequences of calls to tools such as click, type, and scroll on a browser. Current implementations follow a sequential fetch-screenshot-execute loop where each iteration requires an LLM call, resulting in high latency and frequent errors from incorrect tool use. We present agent just-in-time (JIT) compilation, an alternative that compiles task descriptions directly into executable code that is free to include LLM calls, tool calls, and parallelization. Our approach comprises three components: (1) JIT-Planner, which generates multiple code plans, validates each against tool specifications, and selects the minimum-cost candidate; (2) JIT-Scheduler, which explores parallelization strategies via Monte Carlo cost estimation from learned latency distributions; and (3) an invariant-enforcing tool protocol specifying precondition and postcondition state requirements that reduce the rate of generating plans with incorrect tool use. Across 5 web applications, JIT-Planner achieves $10.4\times$ speedup and $+28\%$ accuracy over Browser-Use, while JIT-Scheduler achieves $2.4\times$ speedup and $+9\%$ accuracy over OpenAI CUA.