Unified CLI workspace
Keep Claude Code, Codex CLI, Gemini CLI, other CLI tools, and plain shells in one sidebar. Group projects into folders and name sessions by role.
CLI Manager
CLI Manager is a macOS app for organizing Claude Code, Codex CLI, Gemini CLI, other CLI tools, and terminal sessions in one workspace for parallel work and Git review.
Overview
CLI Manager is a macOS desktop application for organizing multiple command-line AI agents and terminal sessions in one workspace. It supports Claude Code, Codex CLI, Gemini CLI, other CLI tools, apps, and plain shells, with project folders and named sessions for keeping related work easy to find.
The app combines terminal-based agent management with Git workflows and review tools. Users can create separate worktree workspaces, monitor agent state, inspect diffs, and send line-specific feedback back to an agent’s terminal. An optional AI Control API also allows an agent or script to operate visible sessions locally, while a person can observe, type, disconnect the AI, or take over.
Features
Worktrees and source control
Create Git worktrees as separate workspaces, run agents on different branches, and use the built-in panel to stage, commit, push, browse history, and compare changes.
In-app diff review
Open workspace diffs, select lines, and send comments with file and line numbers back to an agent’s terminal. Worktree comparisons include new files and compare against the forked branch.
Agent status and usage awareness
Use official Claude Code and Codex hooks for session status, distinguishing permission prompts from completed work. Usage alerts indicate approaching five-hour or weekly limits.
Watchable AI control
Enable the local AI Control API to let Claude Code, Codex, or scripts open sessions, run templates, send prompts, read screen output, wait for idle, and release or close sessions through MCP or REST.
Session memos
Attach a memo pad to each session. Memos save as you type, persist across restarts, and can be read by an AI working through the Control API.
Use cases
Managing parallel agent work
Keep several agent sessions for the same codebase in one project folder, name them by role, and switch between them without closing sessions that are still running.
Isolated branch development and review
Create a worktree for an isolated branch, assign an agent to it, then inspect the resulting diff and return file- and line-specific comments before merging.
Supervised agent automation
Let Claude Code, Codex, or a script launch a visible terminal session, send a task, wait for the agent, and read its output while retaining the ability to answer prompts or take control.
Monitoring long-running sessions
Use hooks, status indicators, and usage alerts to distinguish active work, permission requests, and completed sessions when several agents are running.
Pros and Cons
Pros
- Supports multiple CLI agents and ordinary shells in one sidebar.
- Includes Git worktrees, source control actions, diff review, and agent feedback in the desktop workflow.
- AI-controlled sessions remain visible and can be watched, interrupted, or reclaimed by the user.
- Free and open source under the MIT license since version 1.5.0.
- Processes data locally; the AI Control API binds to localhost and uses bearer-token protection.
Cons
- Currently macOS-only; Windows is planned and Linux is not on the roadmap.
- The AI Control API is disabled by default and requires local setup, including a token-protected connection.
- Some API-readable terminal output can be partial when the API is enabled after a session is already running.
FAQ
Which platforms does CLI Manager support?
CLI Manager is currently available for macOS, with downloads for Apple Silicon (ARM64) and Intel (x64). Windows is planned, while Linux is not currently on the roadmap.
Which CLI tools can I use with CLI Manager?
You can manage Claude Code, Codex CLI, Gemini CLI, and plain shells. The product also says that other CLI tools and apps can be added.
Is CLI Manager free to use?
Yes. CLI Manager is free and open source under the MIT license since version 1.5.0. The site does not list paid tiers or usage charges.
How does the AI Control API work?
Enable AI Control API in Settings > Agents. It provides a localhost-only, bearer-token-protected MCP and REST API that can open sessions, send prompts, read terminal output, and wait for completion. You can disconnect the AI or switch the API off to end access.
Can I watch or take over an AI-controlled terminal session?
Yes. Sessions opened through the AI Control API appear as ordinary terminals in the sidebar, with an AI-connected indicator. You can read the screen, type into the session, take it over, or disconnect the AI.
Quick Facts
- Category
- Developer Tool
- Platform
- macOS; Apple Silicon and Intel builds
- Primary users
- Developers working with CLI-based AI agents and terminal workflows
- Supported tools
- Claude Code, Codex CLI, Gemini CLI, other CLI tools, apps, and plain shells
- License
- MIT open source since version 1.5.0
- AI Control API
- Optional localhost-only MCP and REST API; disabled by default
CLI Manager Alternatives
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads in Firecracker micro-VMs with private networking and SDK, CLI, or MCP control.
GitBot
GitBot is a local browser workspace for reusable coding bots powered by Claude Code, Codex, or OpenCode, with persistent threads and configurable permissions.
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repos and verifies changes with compilers, debuggers, sanitizers, and tests.
hob
hob is an independent workspace for coding agents, with local control over sessions, terminals, history, routing, and follow-up work.
Manta AI
Manta AI is an autonomous web app testing tool that maps app behavior, catches regressions, and generates tests from a URL, no scripts or selectors needed.
SonOf
SonOf connects to your repo and PM tool, audits your codebase, and turns approved work into shipped tickets with senior engineering review.