Built for Claude Code, Codex & more

The terminal built for AI agents

Split panes, file explorer, AI agent monitoring, and your entire command library — all in one window.

Built-in terminal
AI agent monitoring
Command snippets
CliBook
2 panes
🦀 Claude Code 4h 23m opus-4-6 134 tok $0.08
claude --dangerously-skip-permissions
🦀

Claude Code v2.1.71

Opus 4.6 · Claude Max

~ via v23.11.0 on 🐙 (ap-northeast-1)
npm run build

vite v6.2.0 building for production...

✓ 127 modules transformed.

dist/index.html  0.46 kB

dist/assets/index.js  284.32 kB

Everything you need in one terminal

Split panes, file explorer, command library, and AI agent monitoring — all built in. Available in 8 languages.

Terminal
🔥 4 panes
🦀 Claude Code 4h 23m opus-4-6 134 tok $0.08
claude --dangerously-skip-permissions
🦀

Claude Code v2.1.71

Opus 4.6 · Claude Max

~ via v23.11.0 on 🐙 (ap-northeast-1)
~ via v23.11.0 on 🐙 (ap-northeast-1)
~ via v23.11.0 on 🐙 (ap-northeast-1)

Built-in Terminal

A real terminal lives inside CliBook. Run commands, drag your snippets right in — no more switching windows.

Split Panes

Split into as many panes as you need. Drag to rearrange, resize on the fly — your workspace, your layout.

File Explorer

Navigate your projects visually. Bookmark frequent folders, and drag any file straight into the terminal.

AI Agent Monitor

Running Claude Code or Codex? CliBook detects them automatically, notifies you when they're idle, and tracks usage over time.

Made for the AI agent era

CliBook auto-detects AI coding agents and gives you full visibility into their work.

Auto Detection

Automatically detects Claude Code, Codex, and other AI agents running in your terminal. Zero configuration required.

Usage Dashboard

Track token usage, model costs, and session duration in real time. Know exactly what your AI agents are spending.

Idle Notifications

Get notified the moment an agent finishes or goes idle. Never miss a prompt waiting for your input.

New in v1.10.0

Orchestrate AI agents like a team

Assign specialized roles to each agent — Coder, Reviewer, QA — and let them work in parallel. One AI doing everything produces mediocre results. A team of focused agents delivers production-quality output.

How it works

1

Create a Channel layout

Click + → Channel, choose your working directory and number of workers. CliBook handles the rest.

2

Give the supervisor a task

Describe what you need. The supervisor automatically breaks it down and dispatches subtasks to workers.

3

Watch them work in parallel

Each worker executes independently in its own pane. Monitor progress in the Channel Panel.

New in v1.14.0

Define roles for every agent

Research shows that splitting work across specialized AI agents produces dramatically better results than a single agent doing everything. CliBook lets you assign each worker a specific role — Coder, Reviewer, QA Tester — so they focus on what they do best.

  • Built-in templates — Feature Dev, Code Review, Bug Fix, Refactor. One click to set up the right team.
  • Custom roles — Write your own role descriptions or point to a .md file. Save as reusable templates.
  • Real-time task tracking — See what each agent is working on in the Channel Panel.
Feature Dev Template
S
Project Manager
Break down specs & coordinate
W
Coder
Implement features
W
Reviewer
Find bugs & security issues
W
QA Tester
Run tests & verify edge cases

Multi-Agent

Supervisor dispatches tasks to workers. True parallel execution across separate Claude Code sessions.

One-Click Setup

Prerequisite checks, automatic MCP installation, and guided setup — ready in seconds.

Live Monitoring

Channel Panel shows agent status, heartbeats, and message flow. Click any agent to jump to its terminal.

Multi-Channel

Enable Discord and Telegram simultaneously — send tasks to your supervisor from any chat platform while workers execute in parallel.

+

Send tasks from Discord or Telegram

Enable multiple channels at once — your supervisor receives messages from Discord and Telegram simultaneously, then dispatches work to parallel workers. Control your AI team from anywhere.

Spec-Driven Development · Beta

Write the spec, let AI verify it

Define what you want in bullet points. Send it to your agent. Then auto-verify whether the implementation actually matches — with diffs and evidence.

Terminal
🦀 Claude Code
🦀 Claude Code 2h 15m opus-4-6 $12.40
~/projects/my-app
claude --dangerously-skip-permissions
🦀

Claude Code v2.1.85

Opus 4.6 · Claude Max

~/projects/my-app

I'll implement the user authentication flow based on

the spec. Let me start with the login form...

Writing src/components/LoginForm.tsx

Writing src/lib/auth.ts

Writing src/middleware/session.ts

|
User Auth Flow 92/100
×
Verified 2m ago › Spec

Login form validates email format before submission

Found regex validation in LoginForm.tsx:42

diff

Password requires min 8 chars with mixed case

diff

Error message should not leak whether a user exists

Messages differ: "Invalid email" vs "Wrong password"

diff

Session token in httpOnly cookie, not localStorage

diff

Rate limiting on login endpoint (max 5 attempts/min)

No rate limiter found in auth routes

diff

CSRF token on all state-changing requests

diff
Fix 2 unresolved

Write specs in bullet points

Describe what you want in plain language. Use templates to get started, or let AI optimize your spec.

Auto-verify with evidence

One click to verify. AI compares git diffs against your spec — shows exactly what's met, partial, or missing.

Fix unresolved items

Found gaps? Click "Fix unresolved" to send instructions back to the agent. Iterate until every spec item passes.

Your snippets, always within reach

Save, organize, and instantly access commands with a global hotkey. Need a new one? Type @ai and describe what you need — AI generates it.

Terminal — zsh
~/projects git status
On branch main
Your branch is up to date with 'origin/main'.
~/projects _
⌘ ⇧ P
|

List all containers

$docker ps -a

docker

Start services in background

$docker compose up -d

docker

Clean up unused resources

$docker system prune -af

docker

Follow container logs

$docker logs -f ${CONTAINER}

dockervariable
Copy ⌘↵ Paste to App
4 commands
Copied to clipboard!

Instant Access

Global hotkey ⌘⇧P launches the palette instantly. Search thousands of commands in milliseconds.

Smart Organization

Folders, tags, and pinned favorites. Keep your commands organized the way you think.

Cloud Sync

Sync your commands across all devices. End-to-end encryption keeps your secrets safe.

Variables & Secrets

Dynamic variables with dropdown options. Sensitive data is masked and encrypted.

AI-Powered

Type @ai to generate commands from plain language. AI explains and optimizes existing snippets too.

Claude Code Integration

Meet your Pixel Buddy

Your Claude Code /buddy companion comes to life as a pixel art pet. It roams your terminal, reacts to your coding, and you can dress it up with hats, accessories, and particle effects.

Blob
Legendary · Crown
Cat
Epic · Wizard + Cape
Dragon
Rare · Santa + Wand
Penguin
Uncommon · Top Hat + Bow Tie
Owl
Epic · Sunglasses + Guitar
Axolotl
Legendary · Halo
🎩

40+ Accessories

Hats, eyewear, capes, held items — customize your companion's look with a live preview.

12 Particle Effects

Sparkles, hearts, flames, cherry blossoms, rainbow — add animated flair around your buddy.

🔄

Auto-Sync with Claude

Detects your /buddy from Claude Code — same species, same rarity, pixel art upgrade.

New in v1.5.0

Automate anything with scheduled commands

Set up cron-based schedules to run commands automatically. Health checks, backups, AI agent tasks — on your schedule.

Schedules
Schedules
Daily Git Report

At 01:00 PM

Next: 27m

Shell Error test

Every minute

Health Check

Every 5 minutes

Next: 2m

Daily Git Report

Active

claude -p "git log --since='24 hours ago' --pretty=format:'%h|%an|%s|%cd' --date=short

Review the commits and generate a summary report in markdown format:

## Daily Git Summary

### Stats

  - Total commits

  - Contributors

### Changes

..." --output-format text

At 01:00 PM Dir: ~/Documents/Source/myproject Timeout: 300s

Execution History

2026/3/16 01:00:00 PM 37.0s success exit: 0
2026/3/15 01:00:00 PM 36.5s success exit: 0
2026/3/14 01:00:00 PM 35.6s success exit: 0

Cron Scheduling

Standard cron expressions with human-readable previews. Quick presets for common intervals — every 5 min, hourly, daily, weekly.

AI Agent Templates

One-click templates for Claude Code, Codex CLI, Gemini CLI, Aider, Goose, and OpenCode. Schedule AI tasks in seconds.

Execution History

Full stdout/stderr output, exit codes, duration, and retry attempts for every run. Debug failures instantly.

MCP Server

Let AI agents manage schedules via natural language. Works with Claude Code, Cursor, Windsurf, and VS Code. Learn more →

Model Context Protocol

AI agents that manage your schedules

CliBook's built-in MCP server lets AI agents create, update, and monitor scheduled commands through natural language. Just tell your agent what you need.

Works with Claude Code, Cursor, Windsurf & VS Code
7 tools: create, list, update, delete, toggle, run, view history
One-click install from Settings > MCP Server
Auto-updates with CliBook — always in sync
Claude Code
Create a schedule to run database backup every night at 2am
🦀 Claude

I'll create that schedule using CliBook's scheduler.

Tool: create_schedule

name: "Database Backup"

command: "pg_dump mydb > backup.sql"

cron: "0 2 * * *"

Schedule created! It will run every day at 2:00 AM.

Ready to level up your terminal?

Download CliBook for free. Your AI-native terminal awaits.

Version 1.15.4 • Requires macOS 11+

Intel & Apple Silicon