At PennApps, UPenn’s largest college hackathon, I worked with a team to build a lip reader. This lip reader used machine learning in Python via a series of Google Cloud APIs to transcribe words seen in a computer’s webcam. In addition, the program also can learn a user’s voice by listening to about 10 seconds of audio. Then, it performs text-to-speech with that user’s voice. On my team, I focused on the frontend design and structure. The deployment of this site can be found at https://speakinto.space.