Skip to main content
About this page · Updated 2026-11-04 00:15 UTC

What this page is: Add Motorical to Claude, ChatGPT or Cursor in one step, then pick a goal (transactional email, newsletters and lists, domains) to see what to approve now, what to add later and what only the dashboard can do.

How far to trust this page: the build checked 7 names, paths, tools, scopes and error codes against the real product (0 do not match). Not yet reviewed claim by claim.

Updated 2026-11-04 00:15 UTCRecently updated pages

Connect your agent

Motorical works from your AI assistant through one connector. You add it once, sign in once in your browser, and approve only what you need today. Everything else can be added later, on request.

The address of the connector is the same in every client:

https://mcp.motorical.com/v1/motorical/mcp

:::note Availability of the newsletter and list tools on this address The newsletter and list (Communications) tools are being added to this address. Until they are, a connection to this address does not list them. Connect the dedicated marketing server for now, https://mcp.motorical.com/v1/motorical_communications/mcp, as a second connector, with the same steps below. This note goes away when the tools are on the main address. :::

1. Add the connector​

Pick your client. You need only one.

In a normal terminal, not inside a Claude session, run:

claude mcp add --transport http motorical https://mcp.motorical.com/v1/motorical/mcp

Then start a new Claude Code session, type /mcp, pick motorical and choose Authenticate. Your browser opens: sign in or create your account, set up two-factor authentication, and press Allow.

Any other assistant that supports remote MCP servers with OAuth works the same way: add the address above as a remote (streamable HTTP) server. The hosted server page lists the other clients and which sign-in methods work.

2. Pick your goal​

Each goal below says what to approve now, what to add later, and what only you can do in the dashboard. Ask your assistant "what is my Motorical connection allowed to do?" at any time: it calls motorical_whoami, which lists the permissions held and the ones that could still be added.

Send transactional email​

Receipts, password resets, notifications sent by your app or by the assistant.

  • Approve now: the first sign-in covers the starting set: account state, domains, sandbox, recipients, sending and reading delivery analytics. Nothing else is needed to send a first test email to yourself and check how it went.
  • Add later: webhooks and Motor Block management are listed only after you add them (see "Add a permission later" below). Webhook management also needs a paid plan.
  • Only the dashboard can: publishing the DNS records that verify your sending domain (your DNS provider), changing plan, viewing or regenerating sending credentials, and changing a Motor Block's authentication method.

Run newsletters and lists​

Marketing email to your own lists: lists and contacts, templates, campaign drafts, reports.

  • Approve now: reading and editing lists, contacts, templates and campaign drafts, plus test emails to you. This never emails your contacts.
  • Add later, on request: scheduling and cancelling a campaign (send:comms.campaign), managing workspaces and sending identities (manage:comms.workspace), and managing campaign webhooks (manage:comms.webhooks). Each is asked on its own line the first time it is needed, and scheduling always shows you the campaign, the audience size, the time and every warning before it goes ahead on your clear yes.
  • Needs first: a plan that includes marketing email (overview) and a sending domain verified on the domains tools. Ask the assistant to run comms_get_readiness; it names the first thing still missing. Plans and prices are read live from the motorical://docs/plans resource, never from a page.
  • Only the dashboard can: uploading a large contact file (Comm, then Lists, then Import Contacts (CSV); a chat import takes at most 1001 rows as text), adding people who opted out in your old tool as suppressions before importing, deleting lists, contacts or campaigns, resubscribing or exporting contacts, making a workspace the default, creating a webhook (its signing secret is shown once, to a person), and changing the plan. An assistant cannot do these even if you ask it to.

Moving from another tool? See migrating from Mailchimp.

Manage domains​

Adding a sending domain and checking its DNS.

  • Approve now: reading your domains and their DNS health.
  • Add later: adding and verifying a domain (manage:domains) when you are ready for it.
  • Only the dashboard (or your DNS provider) can: publishing the DNS records. The assistant gives you the exact records and re-checks them after you have added them. Removing a domain is not available to an assistant.

Add a permission later​

When a tool needs a permission your connection does not hold, the assistant is told which one and why. In the dashboard open Settings, then API, then Connected AI Agents, and choose Allow next to the permission. Nothing has to be re-created.

Approve the permission when it is asked for (or add it in the Motorical dashboard under Settings, API, Connected AI Agents), then start a new chat or reconnect the connector if the tools still do not appear.

If tools are missing​

Work through these in order. Most cases are the first or second.

  1. The connector has the wrong name or address. A connector name is only a label, but a typo in it creates a second, separate connection that holds no permissions. If you see a connector with a misspelled name (for example the marketing connector saved as "motorical markting", missing an "e"), remove it and add the connector again with the address above. In Claude Code, claude mcp list shows the names and claude mcp remove <name> removes one. In other clients, delete the connector in the connectors settings and add it again. Check the address character by character; the name does not matter, the address does.
  2. The permission has not been granted yet. A tool you have not been given access to may not be listed at all. Ask the assistant what the connection is allowed to do, or open Settings, API, Connected AI Agents in the dashboard and check. If it is not there, allow it.
  3. The permission was granted but the tools still do not appear. The tool list is read when a chat starts or when the connector refreshes, so an assistant that is already running may keep its old list. Start a new chat, or reconnect (refresh) the connector, and ask again. Do not create a new account or a new key.
  4. It still fails after reconnecting. Remove the connector, add it again with the address above, and sign in with the same Motorical account you approved the permission on. A different account has different permissions.

This page is the same for every client on purpose. Menu names differ and change; the address, the permission names and the steps above do not.

Last updated on