Thank Hut

Let us give thanks with grateful hearts.

Comment

ThankHut

💡 Inspiration

We all want to thank someone in our life for something, but we don't always remember all of them. Therefore we came up with an idea to make web app, where you can add the name of the person and how they helped you. And on thanks giving day, well we send then an AI generated Thank You card on behalf of you.

💻 What it does

  • Write about the person who helped or supported you.
  • AI generate thank you text will be generated automatically related to your response.
  • Send an Thank You email this Thanks Giving to you loved ones.

⚙️ How we built it

  • ML: Python, TensorFlow
  • Frontend: HTML, CSS, JS
  • Backend: Django
  • Database: CockroachDB
  • Linode: hosting and storage
  • Authentication: Auth0

💾 Use of CockroachDB

  • We have used CockroachDB as a primary database because it is an easy-to-use, open-source and indestructible SQL database.

☁️ Use of Linode

We utilized Linode for its hosting and storage. Linode is one of the top IaaS providers and is incredibly easy to use and the free Linode credit from MLH for us to learn and build on Linode was the cherry on the cake! Linode is fast, flexible, and reliable, and we truly enjoyed using it.

✉️ Most Creative Use of Twilio

Sending the thank you note via email to other people who helped you.

🧠 Challenges we ran into

  • We researched the ways we can use to reduce the model's size and found Quantization Aware Training to be useful and effective and we are glad it worked.
  • Completing the project in the given time frame

🏅 Accomplishments that we're proud of

  • Achieving 95%+ accuracy on ML models.
  • Completing all these features in just two days is another.

📖 What we learned

  • Integrate ML models directly without creating an API with Django
  • collaboration.

🚀 What's next for ThankHut

  • Improving the accuracy of the models.
  • Add more features.
Share this project:

Updates