6sense
Ask an AI agent about 6sense account insights, buying signals and keyword trends.
Draft broadcasts, tag subscribers and read email stats from Claude.
Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.
Paste it into any MCP client. Setup docs
The Kit MCP server, which Kit now calls AI Connectors, links a Kit email-marketing account (formerly ConvertKit) to Claude. Its 83 tools read subscribers, broadcasts and stats, and act on the account: tag and segment subscribers, draft broadcasts and sequences, build landing pages and manage webhooks. Kit says sends and deletes wait for your Confirm click inside Kit.
Verified connector
Listed by Anthropic as a partner connector in its Connectors Directory.
Connection checked by Agentman on .
Anthropic states this reflects the level of review a connector received, not a security audit.
Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.
update_landing_page replaces the content tree; Kit refuses an update that would lose content until allow_content_loss is sent after you agree.Only partly. Kit's help article says free creators can connect an AI client and see the available tools, but running any tool returns an upgrade prompt instead of doing the work. Actually using the connector needs a paid Creator or Pro plan. You can upgrade from Settings, Billing in Kit, and the existing connection then starts working without being set up again.
It should pause first, but your client settings decide how firmly. Kit says every tool carries a read, write or destructive tag that clients use to ask before acting, and that sends, deletes, unsubscribes and webhook firing return a link that only runs when you click Confirm in Kit. If you set your client to accept all edits, other changes can go through without asking, so Kit advises reviewing every proposed write.
Because the connector pages through subscribers and each page is a separate tool call. Kit's help article says it returns subscribers a page at a time to keep the AI client's context from filling up, and that the client, not Kit, asks for confirmation on each call. To reduce prompts, ask for larger page sizes or allow read tools to run without asking, if your client offers that setting.
Revoke access in Kit, not only in Claude. Kit's help article says removing the server from your AI client stops it sending requests, but the access token stays valid until you also revoke it on Kit's MCP settings page. Revoking in Kit ends the connection on Kit's side immediately, and the client gets a connection-lost error on its next call. Reconnecting later runs the sign-in flow again.
No. Kit's security FAQ states that Kit does not train any AI model on Kit data, including email content and de-identified subscriber data. Once you connect a third-party client such as Claude, that provider's own data settings in your account decide whether prompts are used for training. Kit says it logs only a categorical intent and an anonymous session ID per tool call, not prompt or response text.
initialize POST to https://app.kit.com/mcp and RFC 9728 metadata at https://app.kit.com/.well-known/oauth-protected-resource/mcp (2026-11-18) · retrieved 2026-11-18https://convertkit.com/ returns a 301 redirect to https://kit.com/ (2026-11-18) · retrieved 2026-11-18Connect Kit once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.