Helping kids read via speech recognition since 2017!

What it does

A web app that gives you feedback on your reading of a passage of a book or any text, keeping track of your location in the text and labeling the word(s) you spoke as either correct or incorrect.

How we built it

Using Google Speech API we converted speech to text and used a series of approaches typically used in bioinformatics to align the spoken words with the text. Sequence alignment algorithms usually used for DNA sequencing were used to correctly track where the reader was in the passage. Since reading is all about how words are sounded out, if the word spoken either directly matched the word or was a homophone, it was labeled as correct.

Challenges we ran into

Utilizing streaming speech instead of asynchronous or synchronous, where output was delayed.

Accomplishments that we're proud of

What we learned

What's next for Speekly

Multilingual support!

Built With

Share this project: