Agent Relay Desktop - One window. Every agent.

by
Agent Relay Desktop is a control plane for the coding agents you already run. Claude Code, Codex, Gemini, Kimi, and OpenCode work through one interface, driving the CLIs on your machine under your own subscriptions. Every run streams turn by turn, with messages, tool calls, and file changes visible as they happen, and you set how much the agent does on its own. When one hits a rate limit, the conversation continues on another instead of stopping. Everything stays on your machine.

Add a comment

Replies

Best
Maker
📌
Hey Product Hunt I kept hitting the same wall. I'd be deep in a session with Claude Code, it would hit its rate limit, and that was the end of it. Open Codex in another terminal, re-explain everything, start over. I built the CLI version first (agent-relay.dev) to handle that handoff from the terminal. Agent Relay Desktop is the same idea in a native macOS window. What it does: • Runs Claude Code, Codex, Gemini, Kimi, and OpenCode through one interface. It drives the CLIs already installed on your machine, so you stay on your own subscriptions and it never touches those credentials. • Streams every run turn by turn. Messages, tool calls, and file changes as they happen. • Lets you decide how much rope the agent gets: confirm each command, batch approve, or full auto. • Keeps the conversation going when an agent hits its limit by continuing it on another one. Everything is local. Chats and transcripts stay on your machine.