The Mailgun plugin adds Mailgun in the Cursor Marketplace as a set of tools your AI agent can use. Ask for an outcome in plain language, like "check my sending health" or "why did these emails bounce," and the agent pulls the data from the Mailgun API and returns a readable summary. You don't need to know which endpoints to call or assemble the results yourself.
The plugin is for working with your account: checking health, preparing campaigns, and investigating problems. It is not designed for sending large volumes of email.
What's in the plugin
The plugin bundles three things so you can install them together:
- Mailgun MCP server: Lets the agent call the Mailgun API across Send, Validate, Optimize, and Inspect.
- Skills: Ready-made workflows the agent runs when your request matches, such as campaign pre-flight or bounce analysis.
- Commands and rules: Slash commands you can run directly, plus guidance that keeps the agent's behavior consistent.
If you've used the Mailgun MCP server before, this is the same server. The plugin removes the manual setup: no repository clone and no JSON configuration.
Before you begin
- Cursor: A current version of Cursor with access to the Cursor Marketplace.
- Mailgun account: An account with at least one domain.
- Mailgun API key: A key from API Security in the Mailgun dashboard. This is the only setting the plugin needs.
- Product access: Skills that use Validate, Optimize, or Inspect require those products on your account.
Your API key grants access to your Mailgun account. Enter it only in the plugin settings, never in a chat message or a shared file.
Install the plugin
- In Cursor, open the Customize area and go to the plugin Marketplace.
- Search for Mailgun.
- Open the Mailgun listing and click Add to Cursor.
You may also see a Sinch plugin in the Marketplace. It covers Sinch messaging more broadly. For Mailgun email workflows, install the Mailgun plugin.
Add your API key
- Open the installed Mailgun plugin in Cursor.
- Enter your Mailgun API key in the plugin configuration. You can do this in the plugin settings or through chat.
- Save. The Mailgun MCP server is now enabled for your account.
After you save, Cursor masks the key. To change it later, return to the plugin configuration and enter a new key.
Try it
Open a chat and describe what you want. You don't need to name a Skill. The agent recognizes which Skill applies and runs it. You can also run a Skill directly with a slash command.
Examples to start with:
- "What can the Mailgun plugin do?"
- "Check my sending health for example.com over the last 7 days."
- "Why did my bounce rate go up this week?"
- "Run a campaign pre-flight for tomorrow's newsletter to 50,000 recipients."
- "Validate these addresses: a@example.com, b@example.com."
- "Why didn't my email to jane@example.com arrive?"
Depending on the request, Cursor may present the results as a canvas or report. The layout is generated by Cursor and can vary between runs.
Available Skills
The plugin includes 15 Skills. Open any Skill in the plugin to read its full description.
- Address validation: Check whether email addresses are deliverable before you send.
- Bounce rate explanation: Understand what is driving your bounce rate.
- Bounce suppression analysis: Review addresses on your suppression lists and why they were added.
- Campaign preflight: Check domain health, list quality, and readiness before a large send.
- Cross-domain reporting: Compare sending results across your domains.
- Deliverability triage: Find and prioritize the most likely causes of delivery problems.
- Domain onboarding and DNS verification: Confirm a new domain is set up and its DNS records are correct.
- Engagement analysis: Review opens, clicks, and other engagement trends.
- Inbound routing troubleshooting: Diagnose why inbound routes aren't matching or forwarding as expected.
- MCP troubleshooting: Diagnose problems with the plugin or MCP connection.
- Missing email investigation: Trace a specific message that a recipient says they didn't receive.
- Platform operator: Work across many domains, subaccounts, or IPs.
- Post-send analysis: Review how a campaign performed after it went out.
- Sending health check: Get a snapshot of domain status, volume, delivery rate, and bounces.
- Sending template QA: Check that a template renders correctly and fix issues before you send.
Example: prepare for a large campaign
A pre-flight check covers what you would normally verify before a big send:
- Ask the agent to run a campaign pre-flight and give it the sending domain and approximate audience size.
- The agent checks the health of your sending domain and recent delivery results.
- It returns a report with checks, risks, and recommended next steps. It may recommend delaying a large send if it finds issues.
- If the agent needs more information, such as a sample list of recipients, it asks. Provide it and it can validate the addresses next.
- Fix the issues it flagged, then send through your normal process.
A common workflow is to build the email template in another tool or with another agent, then use the plugin to check that it renders correctly and your domain and list are ready.
Safety and limits
- Read-only by default. Skills retrieve and analyze data. Skills that could send email or use credits ask for your confirmation first.
- Not for bulk sending. Sending a large campaign through agent tool calls uses a lot of context and risks an unintended send. Run the send itself through your usual application or API integration.
- Limited write access. The plugin does not cover every write endpoint in the Mailgun API.
- Actions that stay with you. Account creation, API key issuance, DNS changes, billing, and high-risk bulk actions are not done by the plugin.
- One account per key. The plugin works on the account that your API key belongs to.
Use the plugin in Claude
The same MCP server and Skills are available in Claude, but not through a one-click marketplace listing. To install, add the mailgun/mailgun-plugins GitHub repository as a plugin marketplace source in Claude, then install the Mailgun plugin from it. Setup and the API key step are otherwise the same.
Troubleshooting
The agent doesn't use Mailgun. Confirm the plugin is enabled in Cursor and that an API key is saved. If the agent still answers without calling Mailgun, name the task directly, for example "Use the Mailgun plugin to run a sending health check," or run the Skill with its slash command.
Requests fail or return authorization errors. Check that the API key is correct and active in API Security in your Mailgun dashboard. Replace the key in the plugin configuration if needed.
If the issue continues, ask the agent to run MCP troubleshooting, or contact Mailgun Support.