Docodial || Stanford Hackathon


When I was younger, my mom refused to go to the hospital without my dad– she had a hard time understanding what the doctor was saying, and was not comfortable asking questions. Different languages can create a barrier between patient-doctor care. Not being able to understand doctors is a huge problem when the topic of conversation is about someone’s health. Although translators exist, and Google Translate might work, it’s impossible for these artificial translation services to understand the nuances related to medicine.



At Treehacks, Stanford’s annual hackathon, my teammate and I wanted to approach this problem by providing an interface for doctors to communicate with interpreters that securely provides them with the necessary patient data, and Docodial was born.

We wanted the ability to livestream interpreters on one end and connect them to the doctor and patient sitting together on the other side. This would also integrate with patient data that would be legally released to the translator so that they are also able to learn the patient’s history and gain context before helping translate. We also integrated a speech-to-text API so that there would be an attempt at a written record for review later on.




After some time spend determining the ideal platform, we landed on the web due to the accessibility and openness it offered. For the Video and live streaming functionality, we used TokBox and for the speech-to-text we used the Web Speech API. The Web Speech API is set up to listen in the entirety of the call, transcribing everything that both parties say (unless your mic is muted either physically or through TokBox. We provided access to this through a ‘Room’ style interface accessible through a dashboard.


Challenges We Ran Into:

  • Authentication of TokBox API.
  • A slow and painful ideation process. We pivoted ideas multiple times before settling on this, and caused even further time restrictions.
  • Time constraints. The hackathon lasted a span of 36 hours, and we worked for less than half of that after pivoting ideas and taking necessary breaks.


Tools Used:

HTML, Javascript, Sketch, TokBox API




What’s Next?

  • Truly secure patient data access
  • Persistent user accounts
  • Finding committed volunteers & interpreters
  • UX surveys & in-depth testing



We won!

We entered two categories: Best Live Video App using TokBox API & Best User Interface — & won both.


Special thanks to the best teammate & classmate: Aidan Harris-Tyrell!

He worked on the back-end, integrating the API and making it a functional website. I worked on the front-end, creating a mockup in Sketch and then translating it into code.