What it does
Pincast uses Google Speech-to-Text api to transcribe audio recording between multiple speakers to get the enntire conversation in a text format with start and end times marked for each word. Furthermore, the diartization API also marks different speakers and the MeaningCloud topic response API finds key words in the conversation and provides categorization for key words. Utilizing the information provided by the aforementioned APIs and a combination of a dynamic front end, we are able to allow the user to pin different parts of the podcast. These pins provide the user the ability to go back and reflect, discuss, and learn about the different concepts talked about in the podcasts.
How we built it
Python, Google Cloud Speech to Text API, MeaningCloud API, BeautifulSoup, Flask, Express.JS, MongoDB, React, Figma, AfterEffects
Challenges we ran into
We had trouble with the communication between our front-end and back-end Python scripts, as well as React. All of the technology we used here was new to us, so we spent a lot of time getting familiar with the technology.
Accomplishments that we're proud of
We’re really proud of the way we overcame our challenges
What we learned
How to collaborate effectively on developing a full-stack, backend intensive application, using multiple API’s and real-time frontend interactivity in React.
What's next for Pincast
We hope to develop a version of the app on mobile as that’s where a lot of individuals listen to podcasts. Additionally, we hope to improve the community-aspect of the app, allowing for a greater degree of knowledge sharing (sharing pins, the ability to see the pins of other users). Finally, making the UI more sleek and having a large library of podcasts to draw from would also be an important priority