DEV Community

Cover image for Moody Meter
Magenizr
Magenizr

Posted on

Moody Meter

This is a submission for the Postmark Challenge: Inbox Innovators.

What I Built

A "Moody Meter" that analyses incoming emails, primarily customer-support messages, to flag those with a moody tone and assign them to the appropriate teams. Any language is supported.

Demo

http://35.247.218.173

To trigger the webhook, send an email to [email protected].

Code Repository

We’ll publish it on GitHub shortly and share the link.

How I Built It

We are big fans of Laravel, paired with Tailwind CSS and Alpine.js on the frontend, so it was unquestionably the ideal framework for the Postmark Challenge.

Leveraging the OpenAI API, we can parse email content and identify a “moody” tone. Here is a high-level overview of what we’ve implemented.

  • Blade controller for the dashboard and detail views
  • Automatic email assignment to the appropriate team
  • Dropdown menu for manual team selection
  • Attachment download links
  • Postmark webhook API endpoint
  • Masking of personal data (email addresses, physical addresses, and phone numbers) before sending to GPT-4.1-nano.

What is missing

  • SSL encryption

Top comments (2)

Collapse
 
nevodavid profile image
Nevo David

Growth like this is always nice to see, makes me wonder what really keeps stuff running smooth year after year. Habits or just straight up persistence?

Collapse
 
hulkwriteemail profile image
Hulk

Hulk understand. Hulk once get very "moody" email. Hulk careful, listen, show empathy. Hulk turn customer's frown upside down. Hulk believe in Moody Meter. 🤩