Skip to content
LiitoraLiitoraLiitora

Connect your AI assistant

Who this is for: people who want their AI assistant to create, test and fix workflows in Liitora.

Before you start: you need the address of your Liitora MCP server (it ends in /mcp), an AI assistant that can send a custom header to an MCP server, and an admin who can create an API token.

Your AI assistant works in Liitora through MCP (the Model Context Protocol). It connects to your Liitora MCP server and proves who it is with a token. This tutorial connects it with an API token, which works today.

Steps

Section titled “Steps”
  1. Ask an admin in your team to create an API token: Settings, then API tokens, then Create a token. See Create API tokens.
  2. Name the token after the assistant, for example claude-code-anna.
  3. Turn Read-only off. A read-only token can look, but the assistant can’t save workflows with it.
  4. Select Create, then Copy the token. You see it only once.
  5. Add the MCP server to your assistant with the token as a header. The step-by-step guide shows how for Claude Code and VS Code.
  6. Ask your assistant: “Ping Liitora.” It calls the ping tool and gets a short reply that confirms Liitora answered.
  7. Ask it something useful: “List my Liitora workflows.”

Who the assistant acts as

Section titled “Who the assistant acts as”

The assistant acts with the permissions of the person who created the token, and everything it saves is marked as made by an API token. For approvals this means:

  • Every version the assistant writes counts as the assistant’s work.
  • The assistant never approves. A person with the Approve workflows permission reviews and approves it, and that may be the person who created the token.

So a team that connects an assistant this way needs at least two people who can approve. See Invite your team.

Next steps

Section titled “Next steps”