Gradientsys: A Multi-Agent LLM Scheduler with ReAct Orchestration

TL;DR

Gradientsys introduces a multi-agent scheduling framework that leverages LLMs and ReAct orchestration to efficiently coordinate diverse AI agents, improving task success, reducing latency, and lowering costs.

Contribution

It presents a novel multi-agent scheduling system with a typed protocol and dynamic planning, enhancing coordination, transparency, and efficiency over existing frameworks.

Findings

Higher task success rates on GAIA benchmark

Reduced latency and API costs

Enhanced transparency with real-time activity streaming

Abstract

We present Gradientsys, a next-generation multi-agent scheduling framework that coordinates diverse specialized AI agents using a typed Model-Context Protocol (MCP) and a ReAct-based dynamic planning loop. At its core, Gradientsys employs an LLM-powered scheduler for intelligent one-to-many task dispatch, enabling parallel execution of heterogeneous agents such as PDF parsers, web search modules, GUI controllers, and web builders. The framework supports hybrid synchronous/asynchronous execution, respects agent capacity constraints, and incorporates a robust retry-and-replan mechanism to handle failures gracefully. To promote transparency and trust, Gradientsys includes an observability layer streaming real-time agent activity and intermediate reasoning via Server-Sent Events (SSE). We offer an architectural overview and evaluate Gradientsys against existing frameworks in terms of extensibility, scheduling topology, tool reusability, parallelism, and observability. Experiments on the GAIA general-assistant benchmark show that Gradientsys achieves higher task success rates with reduced latency and lower API costs compared to a MinionS-style baseline, demonstrating the strength of its LLM-driven multi-agent orchestration.