Anemoi: A Semi-Centralized Multi-agent System Based on Agent-to-Agent Communication MCP server from Coral Protocol

TL;DR

Anemoi introduces a semi-centralized multi-agent system with direct agent communication, enhancing collaboration, scalability, and performance over traditional planner-dependent MAS designs, demonstrated by superior benchmark results.

Contribution

It presents a novel semi-centralized MAS architecture utilizing agent-to-agent communication, reducing dependency on a central planner and enabling real-time structured collaboration.

Findings

Achieved 52.73% accuracy on GAIA benchmark with small LLM as planner.

Surpassed open-source baseline OWL by +9.09% under same conditions.

Demonstrated improved scalability and collaboration in multi-agent tasks.

Abstract

Recent advances in generalist multi-agent systems (MAS) have largely followed a context-engineering plus centralized paradigm, where a planner agent coordinates multiple worker agents through unidirectional prompt passing. While effective under strong planner models, this design suffers from two critical limitations: (1) strong dependency on the planner's capability, which leads to degraded performance when a smaller LLM powers the planner; and (2) limited inter-agent communication, where collaboration relies on prompt concatenation rather than genuine refinement through structured discussions. To address these challenges, we propose Anemoi, a semi-centralized MAS built on the Agent-to-Agent (A2A) communication MCP server from Coral Protocol. Unlike traditional designs, Anemoi enables structured and direct inter-agent collaboration, allowing all agents to monitor progress, assess results, identify bottlenecks, and propose refinements in real time. This paradigm reduces reliance on a single planner, supports adaptive plan updates, and minimizes redundant context passing, resulting in more scalable execution. Evaluated on the GAIA benchmark, Anemoi achieved 52.73% accuracy with a small LLM (GPT-4.1-mini) as the planner, surpassing the strongest open-source baseline OWL (43.63%) by +9.09% under identical LLM settings. Our implementation is publicly available at https://github.com/Coral-Protocol/Anemoi.