Orchestration tools for running multiple Hermes agents together. Fleet management dashboards, swarm coordinators, and multi-agent delegation systems.
TradingAgents applies a multi-agent framework to financial trading: fundamental analysts assess company financials and performance metrics, while sentiment analysts aggregate news headlines, StockTwits, and Reddit chatter. Selected analysts work in parallel using their own tools, bringing different kinds of analysis to the trading process.
Mission Control provides a self-hosted interface for governing agent registration, task dispatch, memory browsing, and cost tracking. It supports multi-agent workflows through a Web UI, CLI, MCP server, and REST API, with local Node.js installation or Docker deployment.
Agent of Empires is a TUI and web control surface for managing multiple coding agents, including Hermes Agent and other named agent tools. It keeps sessions running in the background with tmux and offers sandboxing plus dashboards with diff review.
This Control Room-first template describes a path from a single Hermes Agent on one VPS to specialist teams and automated workflows. Its level-based architecture uses a task bus for orchestrator-to-specialist delegation and handoffs, and includes skills for VPS provisioning, security auditing, and backup management.
SwarmClaw is for building autonomous agent swarms, with support for 23+ LLM providers, durable agent memory, and MCP tool integration. Visual org charts show how agent teams are delegated, setting it apart as a swarm-oriented option.
Minions offers a single-screen mission control for creating, supervising, and reviewing autonomous Hermes Agent work. Its Kanban board tracks task status and review queues, while live streams show agent reasoning and tool execution; proposed task completions require human verification.
This project coordinates 17 specialized agents across research, planning, implementation, quality, and infrastructure. A strict pipeline architecture and Master Orchestrator handle routing and coordination.
Hermes Agent Team is a local web system for collaboration between Leader and Specialist agents, managing an automated task lifecycle. It uses MCP and ACP for local tool and communication handling, with a real-time web UI for agent dialogues and Kanban task flows.
Hermes Agent Metaharness automates outer-loop optimization of benchmark harness code. It runs TBLite and TerminalBench-2 evaluations and provides structured candidate mutation and frontier tracking.
Gladiator stages two competing AI companies, orchestrating nine autonomous agents using Paperclip and Hermes Agent for task execution and learning. A real-time dashboard displays agent tasks, code, and scoreboard metrics.
HermesAgent-MultiModel implements a local Mixture-of-Agents stack by querying models in parallel and synthesizing their outputs, using Hermes for task routing and tool-calling. It is designed for consumer GPUs, with stated optimization for RTX 4080/5080 hardware and no cloud dependencies.