ReevesAgents

FAQ

Does ReevesAgents store my API keys?

No. It starts each CLI with the login that CLI already has, so your keys stay inside it. The only state ReevesAgents keeps is a few JSON files under ~/.reeves, and anything that looks like a key is redacted before it is written.

Does my code leave my machine?

Each CLI talks to its own provider exactly as it would without ReevesAgents, with nothing proxied in between. The Web UI answers on 127.0.0.1 only. To open it from another machine, forward the port over SSH.

Which CLIs does it support?

Claude Code, Codex, OpenCode, Hermes, Kimi, DeepSeek, Pi, Qwen and Aider. It can launch any of them that are installed and signed in, and reevesagents doctor lists the ones it found.

What happens to my agents when I close the UI?

They keep working. Every agent lives in its own tmux window, so you can close everything, come back after lunch and find each one where you left it.

Do I have to let one agent control the others?

Only if you want that. A plain run is just agents working side by side. reevesagents attach <cli> gives one CLI the tools to spawn, read and steer the rest, detach takes them away, and MCP spawning enforces the max_agents limit per run.

Does it run on Windows?

Yes, inside WSL, since the full product is built on tmux. For agent control alone, the separate reevesagents-win package runs natively on Windows as an MCP server.

An Eschatia Labs project. Built by Mert Kaya.