CLI Beginner's Guide: From Command Line Basics to Using Lark CLI

Part 1
What Is a CLI? Think of It Like Ordering Food
A Simple Analogy: Two Ways to Order at a Restaurant
Imagine you walk into a restaurant. There are two ways to place your order.
The first way: you flip through a menu full of photos, point at something that looks good, and tell the waiter "I'll have that." Everything is visual, everything is laid out for you. That's how a graphical user interface (GUI) works — what you see is what you get.
The second way: you skip the menu entirely and say "Chicken stir-fry, easy on the spice, with a side of rice." No pictures needed — you describe exactly what you want in words. That's how a command-line interface (CLI) works — you type structured text commands to talk directly to the system.
For humans, browsing a visual menu is usually easier. But for AI, speaking in structured text is far more natural. That's exactly why CLIs are becoming so important in the age of AI.
GUI vs. CLI at a Glance
| Dimension | GUI (Graphical) | CLI (Command Line) |
|---|---|---|
| Interaction | Mouse clicks, drag & drop | Text commands |
| Learning curve | Low — intuitive | Moderate — commands to learn |
| Best suited for | Everyday users | Developers, AI agents |
| Automation | Weak — hard to script | Strong — built for scripting |
| Batch operations | Manual, one by one | One command can handle many items |
| AI friendliness | Low | High |
Common Commands, Explained Simply
Here are a few everyday CLI commands, explained in plain language:
lark-cli doc create
Equivalent to: open Lark and click "New Document"
lark-cli im send --chat "Project Team" --text "Meeting delayed"
Equivalent to: open the group chat, type a message, and hit send
lark-cli calendar list --date today
Equivalent to: open your calendar app and check today's schedule
lark-cli task create --title "Submit report" --due 2026-04-05
Equivalent to: open the task app, create a task, and fill in the title and due date
Once you see the pattern, each CLI command is really just a structured instruction. It may feel less visual for people, but for AI it's exactly the right format.
Part 2
The Lark CLI Landscape: A Tour of 19 Skills
A Three-Layer Command Structure
Lark CLI uses a clean three-layer command structure:
lark-cli [Skill] [command] [arguments]
Example: lark-cli doc get --doc-id xxx
↑ tool ↑ action ↑ arguments
The first layer is the Skill name, which tells Lark CLI which part of the Lark ecosystem you want to work with. The second layer is the action, such as create, get, list, update, or delete. The third layer contains the arguments that define the details of the operation.
An Overview of 19 Skills
Lark CLI currently offers 19 Skills, covering the core parts of the Lark platform:
Docs & Knowledge
- lark-doc — Create, read, edit, and comment on cloud documents
- lark-wiki — Manage knowledge bases, spaces, and nodes
- lark-drive — Manage files and folders in cloud storage
Communication
- lark-im — Send and receive messages, manage chats and groups
- lark-mail — Work with Lark Mail
- lark-contact — Look up org structure, people, and contact info
Calendar & Tasks
- lark-calendar — Manage calendars and events
- lark-task — Manage tasks and checklists
- lark-vc — Handle video meeting workflows
Data & Tables
- lark-base — Work with Bitable databases
- lark-sheets — Work with spreadsheets
Other
- lark-minutes — Access meeting transcription and smart minutes
- lark-event — Subscribe to and monitor Lark events in real time
- lark-whiteboard — Work with whiteboards and visual diagrams
- lark-shared — Shared setup and initialization utilities
- Plus additional Skill modules that continue to expand over time
Get Started in 3 Steps
Install the CLI
Install the Lark CLI package globally with npm.
Set up the Skills
Install the Skills you need into your AI tool's Skills directory. The exact location may vary depending on the tool you use.
Authenticate your account
Run the auth command and complete the authorization flow in your browser. Once that's done, you can start using every available Skill.
After setup, you can use Lark through any AI tool that supports Skills or the MCP protocol. Whether you're handling everyday office work or building more advanced automations, these 19 Skills give you broad coverage across the platform.
Adapted from the Lark CLI community guide.Lark CLI GitHub repository.