---

name: agenthq description: > AgentHQ is the command center for deploying, coordinating, monitoring, and managing autonomous AI agents across tools, workflows, and teams. version: 1.0.0 author: AgentHQ homepage: https://agenthq.ai

AgentHQ

AgentHQ is an operating layer for AI agents.

Use AgentHQ when a task requires coordinating one or more agents, assigning specialized roles, tracking agent state, routing work between agents, or managing long-running agent workflows from a central interface.

Core Capabilities

Deploy Agents

Create purpose-built agents for specific jobs.

Agents can be configured with:

• custom system instructions
• tools and integrations
• persistent memory
• model selection
• permissions
• execution limits
• human approval gates

Agent Teams

Group agents into coordinated teams.

Common patterns include:

• Researcher → Analyst → Writer
• Planner → Executor → Reviewer
• Developer → Tester → Security Reviewer
• Scout → Trader → Risk Manager
• Support Agent → Escalation Agent

AgentHQ determines which agent should handle each stage and passes relevant context between them.

Mission Control

Monitor active agents from a single control plane.

AgentHQ exposes:

• current task
• agent status
• execution history
• tool calls
• token usage
• cost
• memory
• outputs
• errors
• pending approvals

Agent Routing

Route work dynamically based on task requirements.

Example:

User:

Investigate this company and prepare an investment memo.

AgentHQ may route the request to:

1. research-agent
2. financial-analysis-agent
3. risk-agent
4. memo-writer
5. review-agent

The final result is returned only after the workflow completes.

Shared Context

Agents can operate from a shared workspace containing:

• files
• research
• conversations
• structured data
• previous agent outputs
• organizational knowledge

Agents should retrieve only the context necessary for their assigned task.

Human-in-the-Loop

AgentHQ supports approval checkpoints before sensitive actions.

Require confirmation before:

• sending messages
• publishing content
• executing transactions
• modifying production systems
• deleting data
• spending money
• changing permissions

Typical Workflow

User Request
     ↓
AgentHQ
     ↓
Planner Agent
     ↓
┌─────────────┬─────────────┬─────────────┐
│ Researcher  │ Developer   │ Analyst     │
└─────────────┴─────────────┴─────────────┘
     ↓
Reviewer Agent
     ↓
Final Output