johannesjo/parallel-code

What it solves

Parallel Code provides a unified interface to run multiple AI coding agents in parallel rather than sequentially. It eliminates the manual overhead of managing separate git branches and worktrees for different AI-driven tasks, preventing conflicts and allowing developers to review and merge only the successful results from multiple concurrent attempts.

How it works

When a task is created, the application automatically handles the git plumbing: it creates a new branch and sets up a git worktree to isolate the agent's workspace in a separate directory. It then symlinks necessary directories like node_modules to maintain efficiency. The AI agent (such as Claude Code, Codex, or Gemini) is then spawned within that isolated worktree. Once the task is complete, the user can review the diffs and merge the branch back into the main branch via the sidebar.

Who it’s for

Developers who use AI coding CLIs and want to increase their throughput by running multiple agents on different features or bug fixes simultaneously without cluttering their main working directory.

Highlights

• Multi-Agent Orchestration: Run several agents across different branches in parallel from a single GUI.
• Git Isolation: Automatic creation of branches and worktrees for every task to ensure zero conflicts.
• CLI Integration: Supports popular AI coding tools like Claude Code, Codex CLI, Gemini CLI, and Copilot CLI.
• Sandboxing: Option to run tasks inside project-specific Dockerfiles for added security and environment isolation.
• Review Tools: Built-in diff viewer with inline comments and a progress tracking panel for agent steps.
• Remote Monitoring: Ability to monitor agent progress from a mobile device via QR code over Wi-Fi or Tailscale.