Skip to content

This release adds 3 notable features for engineering teams evaluating rollout.

Published 18d MCP SaaS Integrations
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

ai ai-tools anthropic claude email-api email-marketing
+7 more
llm mailchimp marketing-automation mcp mcp-server model-context-protocol python

Summary

AI summary

Updates Highlights, What's in this release, and Internal across a mixed release.

Changes in this release

Feature Medium

create_campaign accepts campaign_type and variate_settings_json for A/B testing

create_campaign accepts campaign_type and variate_settings_json for A/B testing

Source: llm_adapter@2026-05-21

Confidence: high

Feature Medium

create_audience creates full audience with contact info and defaults

create_audience creates full audience with contact info and defaults

Source: llm_adapter@2026-05-21

Confidence: high

Feature Medium

delete_audience permanently removes audience with safety guards

delete_audience permanently removes audience with safety guards

Source: llm_adapter@2026-05-21

Confidence: low

Feature Medium

get_campaign_advice provides automated campaign feedback

get_campaign_advice provides automated campaign feedback

Source: llm_adapter@2026-05-21

Confidence: low

Feature Medium

get_campaign_locations returns geographic open breakdowns

get_campaign_locations returns geographic open breakdowns

Source: llm_adapter@2026-05-21

Confidence: low

Feature Medium

get_eepurl_activity offers social sharing stats from short URLs

get_eepurl_activity offers social sharing stats from short URLs

Source: llm_adapter@2026-05-21

Confidence: low

Feature Medium

delete_audience allows permanent audience removal with safety guards

delete_audience allows permanent audience removal with safety guards

Source: granite4.1:30b@2026-05-22-audit

Confidence: low

Refactor Low

PyPI publishing now uses GitHub Actions with OIDC, removing static tokens

PyPI publishing now uses GitHub Actions with OIDC, removing static tokens

Source: granite4.1:30b@2026-05-22-audit

Confidence: low

Full changelog

v0.3.0 — A/B campaigns, audience lifecycle, advanced reports

This release closes three highly-requested gaps in the Mailchimp Marketing API
coverage: A/B testing, full audience CRUD, and post-send report analytics. The
server now exposes 87 tools (up from 82) and continues to ship with
read-only and dry-run safety modes.

Fully backwards-compatible — existing tool signatures are unchanged.

Highlights

A/B (variate) campaigns

create_campaign now accepts a campaign_type argument (defaults to
"regular") and a variate_settings_json payload describing the test.
Mailchimp will send each variant to a sample of recipients, then auto-pick a
winner based on your criteria and send it to the remaining audience.

Example:

create_campaign(
    list_id="abc123",
    subject_line="Spring Sale 20% off",
    campaign_type="variate",
    variate_settings_json='''{
      "winner_criteria": "opens",
      "test_size": 20,
      "wait_time": 1440,
      "subject_lines": [
        "Spring Sale 20% off",
        "Last chance: 20% off Spring"
      ]
    }'''
)

Subject lines, from-names, reply-to addresses, send times, and HTML contents
can all be A/B-tested via the same payload (2 to 8 variants each).

Audience lifecycle (create + delete)

Two new tools fill the most-asked-for gap in audience management:

  • create_audience — full list creation with contact info, campaign defaults,
    and CAN-SPAM permission reminder
  • delete_audience — permanent removal with the standard read-only / dry-run
    guards in front of it

Advanced report tools

Three new read-only report endpoints, useful for client-facing audits and
deeper post-send analysis:

  • get_campaign_advice — surfaces Mailchimp's automated feedback on a sent
    campaign (open-rate diagnosis, content suggestions, etc.)
  • get_campaign_locations — geographic breakdown of opens by country and
    region, great for region-targeted follow-ups
  • get_eepurl_activity — social sharing stats (Twitter / Facebook / referrers)
    from Mailchimp's short-URL share link

What's in this release

Added

  • create_audience, delete_audience (closes #9)
  • get_campaign_advice, get_campaign_locations, get_eepurl_activity
    (closes #13)
  • create_campaign extended with campaign_type and variate_settings_json
    for A/B testing support (closes #7)
  • 10 new tests covering the additions, including the variate failure paths
    (missing settings, malformed JSON)

Changed

  • Tool count bumped to 87 across README.md, glama.json, and the
    pyproject.toml description
  • Version bumped to 0.3.0

Internal

  • PyPI publishing now runs through GitHub Actions via PyPI Trusted Publishers
    (OIDC) — no static tokens required

Install

Recommended (no install):

uvx mailchimp-mcp

Via pip:

pip install --upgrade mailchimp-mcp

Configure your MCP client to launch uvx mailchimp-mcp with
MAILCHIMP_API_KEY in the environment. See the
README for the
full configuration block.

Safety modes

Both new write tools (create_audience, delete_audience) respect the
existing safety modes:

  • MAILCHIMP_READ_ONLY=true — blocks all writes
  • MAILCHIMP_DRY_RUN=true — returns a preview of the action without calling
    the API

Feedback

Bug reports, feature requests, and PRs are welcome — see
CONTRIBUTING.md.

Full Changelog: https://github.com/damientilman/mailchimp-mcp-server/compare/v0.2.0...v0.3.0

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Track damientilman/mailchimp-mcp-server

Get notified when new releases ship.

Sign up free

About damientilman/mailchimp-mcp-server

Mailchimp Marketing API integration with 53 tools for managing campaigns, audiences, reports, automations, landing pages, e-commerce data, and batch operations.

All releases →

Beta — feedback welcome: [email protected]