Lindo.ai

Your entire agency. One terminal command away.

Create websites, manage clients, run workflows, and check analytics — from the terminal or from an AI agent. JSON output for machines. Pretty tables for humans.

View on npm Read the docs

Compatible agents

Works with any agent that runs shell commands

OpenClaw

OpenClaw

Open-source AI agent for tasks and automation via Telegram.

InstallGitHub
Hermes Agent

Hermes Agent

Autonomous AI agent by Nous Research.

InstallGitHub
OpenCode

OpenCode

Open-source terminal-based AI coding agent.

InstallGitHub
Claude Code

Claude Code

Anthropic's agentic coding tool for the terminal.

InstallGitHub
OpenAI Codex

OpenAI Codex

OpenAI's open-source terminal coding agent.

InstallGitHub

Plus any custom agent built with LangChain, CrewAI, or plain shell scripts. If it can run lindoai, it works.

Tell your agent:

First install lindoai-cli globally and run lindoai skill to install the skill. Then run lindoai --help to learn all available commands. Finally run lindoai login to authenticate and connect to my workspace.

In action

What it looks like

$ lindoai sites create --name "Bright Smile Dental" --description "Family dentistry in Austin, TX"
✓ Website created: web_abc123xyz
  Preview: https://brightsmile.lindo.agency
  Pages: 5 (generating...)

$ lindoai clients create --email "sarah@brightsmile.com" --name "Sarah Johnson"
✓ Client created: cli_xyz789abc
  Magic link: https://app.lindo.ai/magic/...

$ lindoai workspace credits
┌─────────────┬────────────┐
│ balance     │ 2,340      │
│ used        │ 660        │
│ allocated   │ 3,000      │
└─────────────┴────────────┘

Built for agents

The interface AI agents already understand.

AI agents like OpenClaw, Hermes, OpenCode, Claude Code, and Codex operate by executing shell commands. They don't need a GUI, an API client library, or MCP support. They need a CLI that returns structured data they can parse and act on.

The Lindo CLI was designed for this. --format json on any command gives machine-readable output. Non-interactive auth via environment variable. Clear error messages with actionable suggestions. Idempotent operations safe for retry.

Designed for

Developers and AI agents alike

Dual output

Pretty tables for humans, structured JSON for machines. Switch with --format json.

Agent-native

Non-interactive auth, clear errors, idempotent operations. Built for autonomous AI agents.

Scriptable

Pipe to jq, use in bash scripts, run in CI/CD. Standard Unix tool behavior.

Secure auth

Browser login for interactive use. LINDO_API_KEY env var for headless/agent environments.

Idempotent

Safe to retry. Agents can call the same command multiple times without side effects.

Zero dependencies

npx lindoai works without global install. Or npm install -g for permanent access.

Setup

Running in 30 seconds

Install.

npm install -g lindoai — or use npx lindoai for one-off commands without installing.

Login.

lindoai login opens a browser. Or set LINDO_API_KEY for headless environments and AI agents.

Use it.

lindoai sites list, lindoai clients create, lindoai workspace credits — everything is one command.

Frequently Asked Questions

Who is the CLI for?
Two audiences: developers who prefer terminal workflows, and AI agents (like OpenClaw, Hermes, OpenCode, Claude Code, or Codex) that operate by executing shell commands. If you're a developer who lives in the terminal, the CLI is faster than clicking through a dashboard. If you're building an AI agent, the CLI is the simplest way to give it Lindo access.
How is this different from the MCP server?
MCP is for AI assistants that support the Model Context Protocol (Claude, Cursor, ChatGPT). The CLI is for anything that can run shell commands — scripts, cron jobs, CI/CD pipelines, and AI agents that don't support MCP. Use MCP if your AI client supports it; use CLI for everything else.
Can I use the CLI in CI/CD pipelines?
Yes. Set LINDO_API_KEY as an environment variable, use --format json for machine-readable output, and pipe results to jq or your processing tool. The CLI never prompts for input when an API key is set — it's fully non-interactive.
What AI agents work with the CLI?
Any agent that can execute shell commands: OpenClaw, Hermes, OpenCode, Claude Code, Codex, custom agents built with LangChain or CrewAI, or even a simple script that calls lindoai commands. The JSON output mode makes it easy for agents to parse results and make decisions.
Is there a Docker image?
Not officially, but since it's an npm package, you can install it in any Node.js Docker image with npm install -g lindoai. Set LINDO_API_KEY in your container environment and you're ready to go.
Get started

npm install -g lindoai

One install. Your entire workspace at your fingertips — or your agent's.

View on npm
Book a demoTalk to the team