Kit MCP server for your AI

One field, labeled API key, and it takes a Kit V4 API key (Kit was formerly ConvertKit). You paste it once and the MCP server is ready, with no OAuth popups and no token refresh to babysit. The key belongs to a single Kit account, so the connection is locked to that account.

Subscribers, including the engagement filter and a single subscriber's tags and stats; tags, with tag and untag by email address or id; custom fields; forms; sequences (courses) and their sequence emails; broadcasts (newsletters) with stats and link clicks; segments; email templates; posts; snippets; account stats and creator profile; and webhooks. Anything outside those 13 API modules is not exposed.

62 operations across 13 API modules, split 31 read and 31 write. 35 load into your AI client by default and the rest stay searchable, so a niche call may not appear in the tool list until the agent searches for it. That default keeps the tool list short enough for a client to reason over.

Email address works. Tagging and untagging accept either the email address or the subscriber id, and adding someone to a sequence or a form has an email-address route as well. That matters in practice because the list you paste into an AI chat is usually a column of emails, not Kit ids.

Webhook events cover a subscriber confirming their email, unsubscribing, hard bouncing, or filing a spam complaint; signups through a form or landing page; a sequence starting or completing; a tag being added or removed; a purchase recorded against a subscriber; and custom fields being created, deleted, or having a value change. There is also a polled check that catches a new subscriber added by any route. The server can create and list webhooks in the account too.

Yes, with read-only access, which applies on the Free plan and to teammates: it can only list and report, which suits broadcast stats and growth numbers. With full access, destructive calls like unsubscribing a subscriber, deleting a broadcast, or deleting a sequence run as soon as the AI calls them, so tick Kit on the consent screen only for clients that should make them.

No. PopMCP hosts it and gives you a Streamable HTTP MCP URL. There is nothing to deploy, no callback URL to register, and no infrastructure to keep running or patch.

Any client that supports a remote Streamable HTTP MCP server: Claude Desktop, Claude Web, Claude Code, ChatGPT custom connectors, Codex and the Codex CLI, OpenCode and Z.ai's ZCode, as well as editors such as Cursor, Windsurf and Zed. Your client signs in to PopMCP with OAuth; the provider credential never leaves the server.

Connect Kit in PopMCP first, then run claude mcp add --transport http popmcp https://app.popmcp.com/mcp in a terminal. Inside Claude Code, run /mcp and sign in to PopMCP in the browser window it opens, ticking Kit on the consent screen. The Kit tools are then available in that session, and no Kit credential is stored on your machine.

Loading 62 tool definitions into every conversation burns context before the work starts. PopMCP loads a curated 35-tool set covering the everyday workflows and keeps the rest searchable, so your AI pulls a specialist tool only when a task needs it. Every tool stays searchable, and on Scale the owner can load the full catalog for a teammate's connection.

A tool is something your AI calls, and there are 62 here. A trigger is something PopMCP watches for you: 21 Kit events that can start work on the PopMCP side. Triggers never enter your AI client's context, and none of them ask you to paste a callback URL into Kit.