Inspiration
Spell'd was created to help bridge gaps in English learning by being a high-quality spelling tutor.
What it does
Spell'd does two things: it tests users using TTS and asks them to spell the word. It then identifies if the answer was correct or what error the user made. This is repeated 9 times, breaking learning into short mini-sessions, allowing the user to reflect on errors. Spell'd records statistics for every answer, identifying average speed, overall accuracy, trending accuracy, and providing data on the user's most common errors.
How we built it
Spell'd was built in three distinct parts: a front-end, a back-end, and a database. The front-end is built on Python, HTML/CSS, FastAPI, and Streamlit. It provides a clean user experience and displays Spell'd branding and style. The back-end is built on Python, FastAPI, Pydantic, and SQLite. It effectively communicates between the front-end and the database, communicating statistics to the database and updating the front-end. The database utilizes Python and SQLite. It is a lightweight, cost-free solution that allows statistics to be captured and logged in real-time, completely offline.
Challenges we ran into
By trisecting the workload, we made it hard to work in parallel because each part relies on the architecture of the other parts. This made it hard to get started and to test each section. Additionally, we were unable to add the TTS to the questions because they require an API that costs money. Given funding, however, it is a trivial implementation as the infrastructure is in place to support it.
Accomplishments that we're proud of
We are proud of our product. It is, given the trivial implementation of TTS, customer-ready. The product was well thought out and addresses a real problem. This was also, for all of us, our first hackathon, and we are especially proud of our time management.
What we learned
All of us gained crucial skills in developing modular code (so that we could add each part to each other later). Additionally, we all gained skills in using the respective languages, APIs and features.
What's next for Spell'd
As it stands, Spell'd is a nearly complete project. We can easily add the TTS API, and it is good to go. In the future, Spell'd could be integrated with AI to create a significantly larger database of words and error groupings. It is also feasible that Spell'd be expanded to other languages. Spell'd could generate money through advertisements (as opposed to subscriptions to further democratize access).
Log in or sign up for Devpost to join the conversation.