Missive add-on

The Missive MCP server gives Claude direct access to your shared inbox in Missive: conversations, messages, internal comments, contacts, tasks and analytics reports. You ask a question in Claude Desktop or Claude Code, and Claude looks it up in Missive instead of you scrolling through threads and pasting them into the chat.

It is an add-on, not a standalone server. It runs inside mcp-core, the self-hosted MCP server that you install with Docker on your own machine. Your Missive API token stays in the .env file on your own server.

What can Claude read through the Missive MCP server?

Claude can read almost everything a Missive user sees in the app. The read tools cover:

• Listing conversations, filtered by a contact's email address or domain, and opening a single conversation
• Reading the messages inside a conversation, and fetching one message in full
• Posts, internal comments and drafts on a conversation
• Contacts, contact books and contact groups
• Tasks
• Analytics reports: create a report and fetch the result

In practice that means prompts like "summarise every open conversation with this customer this month", "show the open conversations with acme.com" or "pull the response-time report for last week". Claude picks the right tool and parameters itself.

Which write actions does Missive MCP add?

Write mode on the Missive MCP server adds six actions: send a message, create a task, update a conversation, merge conversations, create a post, and create or update a contact. None of these are available until you switch them on.

Every MCP tool on mcp-core starts read-only. To let Claude send mail or change conversations, add a seat in My Account → MCP Self-hosted with Missive set to read-write. On a read-only seat the write tools are not registered at all, so Claude cannot call them by mistake. Sending also needs MISSIVE_FROM_ADDRESS. Without it, the send tool refuses to send.

How do you set up the Missive MCP server?

You need a running mcp-core install and a Missive API token. You create the token in your Missive settings. Then:

1. Buy the Missive add-on. It is added to your existing license; there is no new download.
2. Add the variables below to the .env next to your docker-compose.yml.
3. Run ./mcp-update.sh to restart with the new config, then docker exec mcp-core npm run refresh-license to pick up the new entitlement right away.

MISSIVE_API_TOKEN=your-missive-token
[email protected]

For read access, MISSIVE_API_TOKEN is the only required variable. After purchase you also get an email that lists these variables. The full walkthrough, including where to find the token, is in the Missive setup guide.

Can different team members get different Missive access?

Yes. Each seat on the Missive MCP server is one person or device with its own token, its own subset of tools and its own read-only or read-write mode. Your support lead can have write access while an analyst only reads conversations and reports. You pick the mode when you add the seat. To change it later, add a new seat with the other mode and revoke the old one.

MCP Self-hosted includes three seats. If more people on your team need Claude with Missive access, add an extra seat for each of them.

Who is this for?

The Missive MCP server is built for teams that run support, sales or operations out of Missive and already use Claude. Typical uses are triaging a busy inbox, preparing a reply from the full thread history, turning a conversation into a task, and pulling numbers from analytics without opening the reports screen.

Missive pairs well with the Brain MCP server: Claude reads the customer thread in Missive and writes the decision or agreement into your notes vault, so the context is still there next week. For a non-technical overview of the whole flow, read how to connect Claude to Missive.