SpeakOnFleek

Recording screen
List of previous recordings

Inspiration

When you talk to your rubber duckies on your desk, you somehow always find the answer. We used that idea of brainstorming and made a notes app based off of normal speech. By just talking, words and images are generated relating to what you were talking about. By being able to visualize and hear what you are talking about, brainstorming is a lot easier.

What it does

Our app takes in audio through a microphone controlled by a button and then converts that into text. This text is then scanned for keywords in the sentences or phrases and then an image of them is displayed in order to help brainstorm. The purpose of this is to make an easy way to brainstorm or take notes while talking out loud through your problems.

How we built it

We used a python backend and flutter frontend to generate photos and videos relating to text generated from speech to text. The text generated from the text to speech portion of the app is put into a firestore database so the server can use and manipulate that data later. After that data was processed by the server, a list of keywords along with a list of links which are used later to show the pictures and keywords from the original text.

Challenges we ran into

We found a library to use in flutter for the speech recognition code, however the code we found was out of date. By scanning through the API itself, I was able to fix the code to be in the most recent version of swift.

Another challenge we faced, was connecting Machine Learning code with Flutter.