swe-mux runs many coding agents side by side in one terminal multiplexer with live per-session status and phone access over Tailscale.
Yes, swe-mux is open source under the Apache-2.0 license.
swe-mux is free to use.
Yes, swe-mux can be self-hosted (the source is available under the Apache-2.0 license).
swe-mux is a terminal multiplexer that runs multiple coding agents (Claude Code, Codex) side by side with live per-session status, reachable from a phone over Tailscale. Open source: yes (Apache-2.0); self-hostable; pricing free. It is Python 3.12+, ships a PyPI wheel, and has 6 stars.
swe-mux sits on top of your existing agent CLIs rather than replacing them: you launch Claude Code, Codex, or other agents inside it, and it tracks which session is working, which is idle, and which is waiting on input. Around the sessions it bundles the operator chores that pile up when you run agents in parallel — terminal management, git review, file management, and local previews — in one place. Because it exposes the multiplexer over Tailscale, you can check on and steer long-running agents from a phone instead of SSH-ing into a desktop. Open source: yes (Apache-2.0); self-hostable; pricing free.
The teams shipping GTM tooling — internal apps, integrations, outbound automations — increasingly do it by running several coding agents at once. The bottleneck stops being any single agent and becomes operator attention: which of six sessions is stuck, which finished, which needs a decision. swe-mux is a control surface for exactly that fan-out — one view of every agent's state, plus git and preview in the same window, reachable from a phone so a stalled run does not wait until you are back at your desk. Open source: yes (Apache-2.0); self-hostable; pricing free. The honest read: it is early (6 stars) and unproven at scale, so treat it as a convenience layer for a workflow you already trust, not as a reliability guarantee.