The problem TriMail solves

The idea of TriMail with a mix of blockchain and AI seemed fascinating to us, imagine being able to not just get whatever newsletter, but only the one that you are really interested in, with the security and privacy of blockchain. TriMail makes it possible with the help of Particle, Polybase ,Tensorflow and a custom SMTP server using Nodemailer

Challenges we ran into

We encountered several challenges throughout the project, including:

1. Integrating Polybase with the Particle network.
2. Custom SMTP server implementation.
3. Integrating the SMTP server with a decentralized database and AI model.
4. Integrating the CORE API with the SMTP server and AI model.

Integrating Polybase with the Particle Network

• Integrating Polybase with the Particle network posed challenge. We created a custom hook and made API calls to finish the integration. Used polybase dynamic NFT to create a SBT(Soul bound token) of the user, an ERC721 token, to prevent data tampering, making user an NFT.

Custom SMTP Server

• Implemented a custom SMTP server as regular mailing was a new experience for us. We utilized cron jobs and the Nodemailer service. Eventually, we integrated Nodemailer with mailgun for efficient email delivery.

Integrating AI with Frontend

• We had no prior experience with SMTP protocols and had to go through the Nodemailer documentation to create a working and efficient SMTP server using express.js. Additionally, we incorporated a cron service to send emails to each user according to their selected schedule, such as daily, weekly, monthly, or specific days like Monday, Wednesday, and Friday.

Integrating the SMTP Server with a Decentralized Database and AI Model

• Connecting the SMTP server with Polybase and the AI model was essential for proper on-site functioning. We aimed to minimize overhead on the frontend by directly utilizing the database and model with the SMTP server.

Integrating CORE API with SMTP and AI Model

• The CORE REST API itself presented complexities. We extensively studied the documentation to achieve our goals. Integrating it with the SMTP server and AI model posed another challenge. Each email required dynamic inputs and outputs, and with multiple users having different interests, we successfully implemented the integration.