peep is a zero-config terminal dashboard for watching AI coding agents work in real time. It reads local session logs from Claude Code, Codex, and Gemini, then turns them into a live view of your lead agent, sub-agents, project activity, and context usage. No hosted dashboard, no API keys, and a safe demo mode with peep --mock.

Hey Product Hunt, I built peep because once multiple AI coding agents are running, raw logs stop being a usable interface. The data is there, but it gets hard to tell which agent is active, which repo is moving, or whether a sub-agent is blocked, waiting, or done. peep is a zero-config terminal dashboard that reads local session logs from Claude Code, Codex, and Gemini, then turns them into a live view of your lead agent, sub-agents, project activity, and context usage. I wanted something terminal-native, local-first, and fast enough to keep open while working. No hosted dashboard, no API keys, and a safe demo mode with peep --mock. If you try it, I’d love feedback on signal vs noise, missing integrations, and what you want to see when a bigger agent swarm is running.

This is such a useful idea. Does it work with custom agent setups or just the defaults?

peep

Watch Claude Code, Codex, and Gemini agents live

Watch Claude Code, Codex, and Gemini agents live