Connect your AI to Mailchimp

What your AI can do with Mailchimp

Once connected, these tools appear automatically in your AI assistant:

• list_audiencesread-only

  List audiences (lists) of the account.

• list_membersread-only

  List members (subscribers) of an audience.

• add_memberwrite · needs approval

  Add (or subscribe) a contact to an audience. Writes to Mailchimp, so it goes through the confirmation guardrail when enabled.

• list_campaignsread-only

  List campaigns, optionally filtered by status.

• get_campaign_reportread-only

  Get the performance report of a campaign: opens, clicks, bounces, unsubscribes, plus `list_stats` (this audience's own averages) and `industry_stats` (the sector's) — the two figures that tell you whether a result is good, rather than merely what it is.

• create_audiencewrite · needs approval

  Create an audience. Mailchimp demands the details that make a mailing lawful, so they must come from the customer and cannot be invented: a physical postal address shown in every footer, the reminder telling recipients why they are hearing from you, and the default sender identity. Note that Mailchimp bills contacts per audience — the same person in two audiences counts twice.

• add_memberswrite · needs approval

  Add several contacts to an audience in one call. Each contact's `status` is a statement about consent: 'pending' asks them to confirm by email, 'subscribed' asserts they already agreed. It defaults to 'pending' — only say 'subscribed' when the customer confirms they hold that consent.

• get_memberread-only

  Get everything Mailchimp holds on one contact: status and unsubscribe reason, merge fields, tags, language, VIP flag, source, its own average open and click rates, a 1-to-5 rating, location — and the consent trail (signup and opt-in IP and timestamp, marketing permissions). Addressed by email; the contact id is derived here.

• get_member_activityread-only

  Get a contact's recent activity — the campaigns they opened, clicked or bounced on. Narrow it with `activityFilters` to ask only for the events that matter, rather than reading the whole feed to find them.

• search_membersread-only

  Search contacts across every audience by name or email, without knowing which audience they belong to.

• list_segmentsread-only

  List an audience's segments and tags — how the audience is actually cut up for targeting.

• get_segmentread-only

  Get one segment, with how many contacts it currently holds and — for a rule-based segment — the conditions behind it. Read an existing segment before writing one: its `options` show the exact shape a condition takes for each field type.

• create_segmentwrite · needs approval

  Create a segment of an audience, in one of two shapes. Give `emails` for a static segment — a fixed list you fill yourself, which is what a tag is. Give `match` and `conditions` for a rule-based segment, which Mailchimp recalculates on its own as contacts change. A condition's exact shape depends on the field it tests, so copy one from an existing segment read with get_segment rather than guessing.

• update_segmentwrite · needs approval

  Rename a segment or replace its conditions. Mailchimp requires the name on every call, even when only the conditions change — so pass the current name if you are not renaming it. Replacing the conditions replaces them entirely.

• update_segment_memberswrite · needs approval

  Add or remove contacts of a static segment, up to 500 of each per call. Only works on a static segment: a rule-based one has no fixed membership to edit, change its conditions with update_segment instead. Addresses absent from the audience are ignored rather than created.

• search_tagsread-only

  Search an audience's tags by name, or list them all when no name is given. Call this before add_member_tags: Mailchimp creates a tag silently when the name does not match an existing one, so a difference of case or spelling quietly makes a second tag nobody notices.

• add_member_tagswrite · needs approval

  Tag or untag a contact. Each tag carries its own status, so one call can both add and remove. Tags are the everyday way to segment in Mailchimp — a contact added without one is hard to reach again.

• create_campaignwrite · needs approval

  Create a campaign. It is born as a draft and stays one: this connector exposes no way to send or schedule, because that reaches thousands of people and cannot be undone. Give it content with set_campaign_content, then a human sends it from Mailchimp.

• update_campaignwrite · needs approval

  Change a draft campaign's subject, sender, preview text or recipients, without touching its content. Only a campaign that has not been sent can be changed.

• get_campaign_contentread-only

  Get a campaign's content, in HTML and plain text.

• set_campaign_contentwrite · needs approval

  Set a campaign's content — raw HTML, plain text, or a template with its sections filled in. Replaces whatever was there. The campaign remains a draft until a human sends it.

• list_templatesread-only

  List the email templates of the account.

• get_templateread-only

  Get one template — its name, category and when it was last edited.

• create_templatewrite · needs approval

  Create a template from raw HTML. Mailchimp's API only creates Classic templates: the result will not open in the drag-and-drop editor, so the customer will have to come back here — or to raw HTML — to change it. Mark the parts a campaign should fill with mc:edit attributes, otherwise the template has no editable sections.

• update_templatewrite · needs approval

  Replace a template's name and HTML. Despite being a PATCH, Mailchimp requires both every time — there is no way to amend part of a template, so read it first and send back the whole thing. Classic templates only, like create_template.

• list_campaign_reportsread-only

  List the reports of sent campaigns. Bound the period with `sinceSendTime` and `beforeSendTime` to compare one quarter against another — Mailchimp has no comparison endpoint, so this is how a trend gets built.

• get_campaign_open_detailsread-only

  List who opened a campaign, contact by contact — the report to cross with a CRM to find who is paying attention. Automated opens are excluded by default; since mail apps began preloading images, counting them overstates interest.

• get_campaign_click_detailsread-only

  List the links of a campaign and how many contacts clicked each one.

• get_campaign_email_activityread-only

  Replay every recorded action of a campaign contact by contact — opens, clicks and bounces, each with its timestamp. The heaviest report of the connector: prefer get_campaign_open_details or get_campaign_click_details for a plain answer, and use this with `since` to pick up only what has happened since a given moment.

• get_campaign_recipientsread-only

  List who a campaign was sent to, with each contact's delivery status.

• get_campaign_abuse_reportsread-only

  List the spam complaints a campaign drew. Small, and the earliest warning that deliverability is slipping.

Example prompts

Just ask your AI in plain language — it picks the right Mailchimp tool:

• “List audiences (lists) of the account — from Mailchimp, in plain language.”
• “List members (subscribers) of an audience — from Mailchimp, in plain language.”
• “List campaigns, optionally filtered by status — from Mailchimp, in plain language.”
• “Get the performance report of a campaign: opens, clicks, bounces, unsubscribes, plus `list_stats` (this audience's own averages) and `industry_stats` (the sector's) — the two figures that tell you whether a result is good, rather than merely what it is — from Mailchimp, in plain language.”

How it works

1. Create your Proxoria account and add your MCP endpoint to Claude or ChatGPT.
2. Connect Mailchimp with your API key — each user connects their own account, credentials encrypted at rest.
3. Your AI instantly gets the Mailchimp tools. Write actions wait for your approval in the dashboard before anything changes.

Frequently asked questions

Does Mailchimp have an official MCP server?

You don't need one. Proxoria exposes Mailchimp to your AI through the Model Context Protocol, so you can use it from Claude, ChatGPT, Mistral or any MCP-compatible client — no official Mailchimp MCP server required.

What can my AI do with Mailchimp through Proxoria?

It gets 31 tools: list_audiences, list_members, add_member, list_campaigns, get_campaign_report, create_audience, add_members, get_member, get_member_activity, search_members, list_segments, get_segment, create_segment, update_segment, update_segment_members, search_tags, add_member_tags, create_campaign, update_campaign, get_campaign_content, set_campaign_content, list_templates, get_template, create_template, update_template, list_campaign_reports, get_campaign_open_details, get_campaign_click_details, get_campaign_email_activity, get_campaign_recipients, get_campaign_abuse_reports. For example, list audiences (lists) of the account.

Is the Mailchimp connector read-only?

No — it can also make changes. Write actions go through Proxoria's confirmation guardrail by default: your AI proposes them, and you approve each one from your dashboard before anything is written to Mailchimp.

How do I connect Mailchimp to my AI?

Create a Proxoria account, connect Mailchimp with your API key, and add your Proxoria MCP endpoint to Claude or ChatGPT. Your AI instantly gains the Mailchimp tools.