A self-hosted control plane for AI agents. Let agents work across your projects, tools, workflows and infrastructure while keeping context, permissions and execution evidence under your control.
One runtime, three ways in: desktop/mobile browser, your terminal, or any MCP client.
Project context, trusted skills and task-specific tool discovery.
Real PTY, bounded file tools, service controls and explicit approvals.
Durable sessions, local memory, workflow evidence and agent handoffs.
Schedule/webhook triggers, branching, loops, retries, versions and private variables.
Linux, macOS, Windows/WSL2, Android/Termux, iOS/iPadOS client.
ChatGPT, Codex, Claude Code, Cursor, or any compatible MCP client.