Vocowbulary Courses

• 

  Vocowbulary Exercise

• 

  Vocowbulary Answer

Inspiration

We were inspired by close family members that struggle to pronunciate English words as it is not their first language. We hope to build a tool that is beneficial and entertaining in practicing a useful skill.

What it does

Vocowbulary compares the IPA phonemes for a given word and the extracted phonemes for a user audio recording of the same word and compares the phonemes to give suggestions. Additionally, users auto rate themselves which is used for a space repeated algorithm so difficult words are practiced more often.

How we built it

The frontend is built using the React JS framework, the backend was using the Python Flask framework, and the MongoDB database was connected through Node JS. For the audio-phoneme extraction, the Allosaurues python package was utilized, the Words JS API was utilized for a ground truth of phonemes, and the Gemini LLM model was utilized to compare the phoneme sets.

Challenges we ran into

Major challenges we ran into were merge conflicts despite distributing the work as separate as possible and type conversions for certain JavaScript functionalities due to the dynamic type nature of the language.

Accomplishments that we're proud of

Building a full tech stack product with a custom space repetition algorithm connected to an online MongoDB database.

What we learned

We learned how to utilized MongoDB Atlas, different audio features and qualities, and React specialties such as hooks.

What's next for Vocowbulary Courses

Expand the vocabulary to a larger dataset and allow users to input the word they would like to practice. Additionally, it could be expanded to support other languages.

Built With

• allosaurus
• css3
• express.js
• extendedaudioencoder
• flask
• gemini
• html5
• javascript
• linux
• macos
• mongodb
• node.js
• python
• react
• speechify
• typescript
• windows-10
• words