Skip to main content
Use the CLI with an AI coding assistant to manage NeetoAuth in plain language. The neetoauth setup commands add NeetoAuth instructions for your assistant, while --toon keeps command output compact for it to read.

​
Set up your assistant

Run the subcommand for the tool you use: Every subcommand except setup claude writes into the current directory, so run it from the root of the project the assistant works in. The copilot, gemini, and codex commands place their section between <!-- neetoauth:start --> and <!-- neetoauth:end --> markers and replace that section on re-run instead of adding a duplicate. Re-run the command after every CLI upgrade so the instructions match the installed version.
setup claude prints two slash commands to finish the installation inside Claude Code:
After setup, your assistant can check the commands and flags available in your installed CLI version:
The command prints the complete catalog as JSON.

​
Tips for agent workflows

Prefer --toon when the assistant needs to read a list or record, and --quiet when it needs to pipe an identifier into the next command. See Output formats.
  • Authenticate once with neetoauth login; the assistant reuses the stored credentials.
  • When several workspaces are signed in, the assistant must pass --subdomain <name> on each command.
  • Run neetoauth doctor first to confirm the assistant is authenticated and connected.
  • Have the assistant run neetoauth products list before inviting anyone, so it uses product and role names that exist in the workspace.
The CLI acts as the person who signed in, so an assistant can only do what you can do. To give an assistant workspace-wide access instead, connect the MCP server with an API key.