Hey Product Hunt — I’m Tien, the maker of OxiMux. 👋 I built OxiMux because running several coding agents from separate terminals quickly became harder to manage than the work itself. Sessions were scattered, agents could collide in the same checkout, and reviewing what each agent changed meant constantly switching tools. OxiMux puts that workflow into one Rust-native cockpit: • Run Claude Code, Codex, Pi, or any ACP agent side by side • Give every task an isolated Git worktree and branch • Review staged and unstaged changes, word-level diffs, and commit history in the same app • Keep terminals alive across app relaunches • Use your existing agent CLIs and logins—OxiMux does not proxy inference or store credentials It’s free and open source under Apache-2.0, available now for macOS 13+ and Windows x64. It’s pre-1.0 and already being used to build OxiMux itself, so I’m especially interested in honest feedback from developers who run multiple agents or worktrees: where does your current workflow break down, and what would make this useful every day?

OxiMux - Run coding agents in parallel from one native cockpit

Replies