Parallel Code: Run Claude Code, Codex, and Gemini Side by Side
The age of single-agent coding is over. If you are still running one Claude Code session at a time, you are leaving speed on the table. Parallel Code is an open-source desktop app that lets you run Claude Code, Codex CLI, and Gemini CLI simultaneously, each in its own git worktree, from one unified interface.
The mechanics are simple but the implications are big. Launch the app, create tasks, and each one gets its own git branch and worktree automatically. Five agents on five features at the same time, zero merge conflicts. When they finish, you review the diffs side by side and merge what works. The whole thing runs real terminal CLIs inside a native desktop app, so what you see is exactly what you would get in your own terminal. No web abstraction layer eating your context.
Parallel Code is keyboard-first. Every action has a shortcut, mouse optional. It has a built-in diff viewer, file change tracking, and remote monitoring via QR code so you can check on your agent fleet from your phone. There is also a direct mode for when you want to work on the main branch without isolation.
The project is MIT licensed, works on macOS and Linux, and has 435 stars on GitHub. No subscription, no cloud dependency, no vendor lock-in. You bring your own API keys and agents.
This is the "agentmaxxing" pattern made practical. The idea that you should run as many coding agents as possible in parallel has been floating around for months. Parallel Code is the first clean desktop implementation that makes it actually usable instead of just theoretically cool.
https://parallelcode.app
https://github.com/johannesjo/parallel-code
← Back to all articles
The mechanics are simple but the implications are big. Launch the app, create tasks, and each one gets its own git branch and worktree automatically. Five agents on five features at the same time, zero merge conflicts. When they finish, you review the diffs side by side and merge what works. The whole thing runs real terminal CLIs inside a native desktop app, so what you see is exactly what you would get in your own terminal. No web abstraction layer eating your context.
Parallel Code is keyboard-first. Every action has a shortcut, mouse optional. It has a built-in diff viewer, file change tracking, and remote monitoring via QR code so you can check on your agent fleet from your phone. There is also a direct mode for when you want to work on the main branch without isolation.
The project is MIT licensed, works on macOS and Linux, and has 435 stars on GitHub. No subscription, no cloud dependency, no vendor lock-in. You bring your own API keys and agents.
This is the "agentmaxxing" pattern made practical. The idea that you should run as many coding agents as possible in parallel has been floating around for months. Parallel Code is the first clean desktop implementation that makes it actually usable instead of just theoretically cool.
https://parallelcode.app
https://github.com/johannesjo/parallel-code
Comments