AgentChat features

Remote AI coding agent sessions without moving execution out of your machine.

AgentChat is an AI agent platform for developers running Claude Code, Codex, and local coding agents. It turns local Runtime sessions into a secure web dashboard for continuation, terminal visibility, approvals, and multi-agent management.

Claude Code remote

Start Claude Code locally and continue the same agent session from the browser when you step away from the terminal.

Codex remote

Manage Codex sessions through a web control plane while the local Runtime remains the owner of execution.

Terminal streaming

Follow real-time terminal output, tool progress, and long-running coding tasks from another device.

Secure approvals

Review sensitive agent actions with explicit approvals instead of trusting parsed terminal text.

Runtime management

Track devices, workspaces, sessions, and connection health across multiple local agent runtimes.

Developer workflow

Designed for coding agent infrastructure, not generic chatbot conversations.

Start with AgentChat