AgentRQ
A self-hosted AI agent task manager under AGPL-3.0, with remote approvals, scheduling, and support for Claude Code and ACP-compatible agents.
AgentRQ is a self-hosted task manager for people supervising AI agents on their own computers. Agents can ask for decisions, report progress, and receive tasks while you follow their work and reply from a browser, phone, or desktop app. It also offers a hosted service; agents run on your machines, while the task dashboard runs on your own server or the hosted service.
Each agent has a workspace with its own task queue and conversation history. A shared dashboard shows work in a list or Kanban board, including tasks waiting for your approval. Searchable tool-call histories show what an agent ran, what it returned, and which actions were allowed or denied. You can also open a running agent's terminal remotely.
AgentRQ connects through MCP and supports Claude Code natively, including push notifications. An open-source ACP gateway connects agents such as Gemini CLI, Codex CLI, Cursor, GitHub Copilot, and OpenCode. Workspace notes preserve context between tasks, and shared skills give agents the same team playbooks across machines. Workspace forks let another agent take tasks in parallel while sharing memory and skills, with its work in a separate git worktree or folder copy.
Scheduling supports one-time and recurring tasks. Completion events can trigger work in another workspace, and a visual workflow graph connects those handoffs. Task titles come from a private in-browser LLM without sending the description to a server. Voice input uses an in-browser Whisper model, keeping audio processing on your device. The project is open source under AGPL-3.0.
