Summary
A macOS terminal for managing parallel Claude Code and Codex tasks with integrated Git workflows.
About
Farol coordinates multiple coding-agent sessions in a macOS terminal and connects their task status to isolated Git worktrees and common code-review workflows.
Farol is a macOS terminal for running Claude Code and Codex in parallel. Its sidebar indicates whether each agent is working, waiting for input, or finished, while a menu bar view provides status across applications. Each task runs in its own Git worktree and can be started from a prompt, GitHub issue, or Jira ticket. Users can review diffs, commit changes, push branches, and open pull requests in the same window. Farol is built on libghostty and reads Ghostty configuration. It is free, MIT licensed, requires no account, and does not use telemetry.
Key Features
Runs Claude Code and Codex in parallel.
Displays agent states in a sidebar and menu bar.
Creates a separate Git worktree for each task.
Starts tasks from prompts, GitHub issues, or Jira tickets.
Supports diff review, commits, pushes, and pull-request creation in one window.
Real-World Use Cases
Parallel coding-agent task management
Developers can monitor multiple Claude Code and Codex tasks and identify which sessions need input or have completed.
Issue-to-pull-request workflows
Teams can start work from GitHub issues or Jira tickets, isolate changes in worktrees, and proceed through review and pull-request creation in the terminal.
Concurrent feature development
Individual developers can separate simultaneous coding tasks into distinct worktrees to reduce overlap between agent-generated changes.
Business Opportunities
Increase developer throughput on parallel tasks
Problem: Managing several coding-agent sessions can make it difficult to track progress, identify requests for input, and keep changes isolated.
Solution: Use Farol's agent status views and per-task Git worktrees to coordinate concurrent work and manage changes through pull requests.
Estimated ROI: Potentially positive through reduced context switching and clearer task coordination, but the actual return depends on agent usage, task volume, and team workflow.
Execution: Pilot the terminal with developers already using Claude Code or Codex. Select parallel tasks sourced from prompts or issue trackers, then assess task visibility, worktree separation, and completion of review-to-pull-request workflows.
Standardize agent-assisted issue handling
Problem: Moving coding tasks from issue trackers into agent sessions and then back into code review can involve fragmented tools and manual coordination.
Solution: Start tasks from GitHub issues or Jira tickets and manage their isolated code changes, commits, pushes, and pull requests within Farol.
Estimated ROI: Potential savings in workflow coordination and task setup; no quantified return can be established from the available product information.
Execution: Choose a small set of issue types for a workflow trial, document how tasks are launched and reviewed, and evaluate whether the integrated terminal reduces handoffs without changing existing review standards.