For marketers: check, schedule, and send email campaigns from ChatGPT or Claude

Lettr, an email platform for transactional and marketing email, hosts an MCP server, and one sign-in connects ChatGPT or Claude to it. MCP is the open standard that connects AI assistants to outside services. Once connected, an assistant can report how a campaign performed, build the segments and topics for the next one, and schedule or send a draft from the chat. The campaign itself is still designed in Lettr's editor.

Connect ChatGPT or Claude by signing in

The connection needs a Lettr account and the server URL, https://app.lettr.com/mcp. Signing in to Lettr authorizes the assistant, and the free plan includes the MCP server.

In ChatGPT

ChatGPT adds custom MCP servers as plugins, on the web. They aren't available on every plan, and on Business, Enterprise, and Edu workspaces an admin decides who can add one, as OpenAI's guide to custom MCP apps explains.

• Open Plugins, select the plus button, and choose "Add custom MCP server".
• Enter a name, add https://app.lettr.com/mcp as the server URL, and choose OAuth for authentication.
• Confirm the risk warning, select "Create as a plugin", and sign in to Lettr.

In Claude

• Open Customize, then Connectors, and select "Add custom connector".
• Enter https://app.lettr.com/mcp as the server URL and select "Add".
• Sign in to Lettr when Claude asks to connect.

On Claude's Team and Enterprise plans, an Owner adds the connector for the organization first, and each member then connects with their own Lettr account.

How did the campaign perform?

Every campaign carries its stats, so a question like "how did last week's newsletter do?" returns its deliveries, bounces, opens, clicks, and unsubscribes. The assistant can also list one campaign's individual events, such as each open, click, and bounce, and narrow them to one recipient or a date range.

For the account as a whole, Lettr's sending-health check returns delivery trends, recent problems, suppression hints, and links to the matching dashboard pages. The tools reference lists every tool the assistant can call.

Who will the send reach?

The assistant can build the segments, lists, and topics a campaign sends to. A segment is created by describing its conditions in plain language ("contacts in Germany who joined this year"), and conditions can match contact fields, custom properties, topics, and lists. Lettr returns the segment's contact count when it is created, and the assistant can look up the count for any list or segment.

An opt-in topic (a topic contacts only receive after choosing it) also takes one request. Which segment or list a campaign targets is chosen in Lettr's editor, along with the content and sender.

An AI assistant sends everything it gets from Lettr to its provider, contact details included, so check that provider's data terms before asking about individual contacts or their activity.

Schedule the send from the chat

Scheduling a ready draft is one request. A draft is ready once it has a subject, a sender, content, and an audience, all set in Lettr's editor. Asking for a new time reschedules a scheduled campaign directly, and unscheduling returns it to a draft.

A campaign that starts sending goes out to its whole audience right away and cannot be paused or cancelled. Sending and scheduling therefore carry the MCP destructive annotation, which lets ChatGPT or Claude ask for confirmation first, so a misread request can be stopped before anything goes out.

Why else use Lettr?

Lettr has the rest of a campaign program outside the chat. Its editor has synced sections that update every template using them, multilingual templates, and version history. One campaign can go out in several languages, and double opt-in, unsubscribe handling, and a record of each contact's consent changes are built in.

The same account also sends transactional email, so marketing and transactional email share one editor and one bill.

FAQ

Bottom line

The work around a campaign moves into the chat: results, segments, topics, and the schedule. The campaign itself stays in Lettr's editor, and the destructive annotation lets the assistant ask before anything goes out.

Create a free Lettr account to try it on your own campaigns; the free plan includes the MCP server. Once you're in, the setup guide walks through each assistant.