Lookio Docs
Lookio websiteOpen Lookio
Get startedConnect an AI agent

Connect an AI agent

Set up Lookio from Claude Code, Codex, Cursor, ChatGPT or Claude. The CLI when the agent can run commands, with every permission asked at once; the MCP server otherwise.

This page is written for the AI agent. The person copies the setup prompt from the app (Connect your AI agent), pastes it into their agent, and approves access in their browser. The agent never asks for an API key, a password or a code from an email.

1. Choose the route

The agent can…Route
Run shell commands: Claude Code, Codex, Cursor, Gemini CLI, Windsurf, any agent with a terminalThe CLI. Use it whenever it is available.
Only use connectors: Claude.ai, ChatGPT, Claude Desktop without a terminalThe MCP server.

The CLI reaches everything Lookio does, answers in JSON and uploads files straight from the disk. The MCP server is the route for agents that cannot run commands.

Install it

Terminal
npm install --global @lookio/cli

It needs Node.js 20 or later. Without a global install, run every command below as npx @lookio/cli … instead of lookio ….

Sign in with every permission at once

Terminal
lookio auth login --scope account:read,usage:read,resources:read,resources:write,resources:delete,assistants:read,assistants:write,assistants:delete,query,admin,webhooks:write

Ask for all of them now: a missing permission later means signing in again. Your browser opens, you choose the workspace and approve. Lookio grants only what your role allows (admin needs a workspace admin), and that is expected, not an error.

On a machine without a browser, add --device: the CLI prints an address and a code to open on any device.

Check the connection

Terminal
lookio auth status

It shows the workspace, the plan and the permissions granted.

Learn the commands

Terminal
lookio help --json --compact

The whole command surface as JSON. lookio help <command> --json details one command. See CLI commands.

3. The MCP server (fallback)

Add a remote MCP server with this URL, over HTTP:

https://mcp.lookio.app/mcp

The client opens Lookio in the browser: choose the workspace and approve every permission offered, for the same reason as above. Connect an AI client gives the setting for Claude, Cursor, ChatGPT and others.

4. Then

  1. Tell the person what is connected: the workspace and the permissions granted.
  2. Ask what they want to do first: add resources (files, web pages, a whole site), create an assistant that answers from them, or ask a question.
  3. Before deleting anything, or a job that spends many credits (a large site, many questions), say what it does and wait for their go.

The Quickstart and Concepts explain resources, assistants and credits.

Use these docs with your AI tools

An AI agent can read this documentation directly. You do not need an account or an API key. Everything here is public and read-only.

Query these docs via MCP

Recommended

Add this server to Claude, Claude Code, Cursor, Mistral, or any tool that supports MCP. Your agent can then search Lookio Docs documentation and read it in full, instead of answering from memory.

https://lookio.app/docs/mcp
  • searchFind the passages that answer a question.
  • fetchRead one page in full, as Markdown.
  • list_pagesSee every page in this documentation.

Query these docs over HTTP

The same tools also work as plain web requests. Use this for scripts, or for any tool that does not support MCP. There is one endpoint per tool. Arguments go in the query string, and the answer comes back as JSON.

https://lookio.app/docs/api/docs/search?query=custom+domain

Read the OpenAPI description. It is built from the same definitions as the tools, so it always matches what the endpoints do.

Read these docs as Markdown

Add .md to any page URL to get its Markdown source. You can also send the headerAccept: text/markdown to the page URL itself.

To read the whole documentation in one file, open llms-full.txt. For a short index of every page, open llms.txt.