AgentRQ

AgentRQ is a real-time task management platform for AI agents, solving the problem of keeping humans in the loop during autonomous agent operations. It connects agents like Claude Code and ACP-compatible agents to human operators via MCP notification channels, enabling task assignment, approvals, and replies. The platform targets developers, operations teams, and B2B users, particularly those using AI agents in their workflows. It is positioned as a human-in-the-loop control plane, offering isolated workspaces, real-time notifications, and a visual task board, with a free tier and self-hostable open-source option.

Key features

• MCP native integration
• Real-time notifications via SSE
• Visual task board
• Isolated workspaces
• YOLO mode for auto-approval
• Bidirectional file exchange
• Parallel event streams
• Dedicated MCP endpoints