Inspiration
To make a conference call can sometimes be complicated when coworkers don't all speak the same language. We decided to build spoken.tech, a real-time voice to chat language translation that integrates to CISCO Spark.
What it does
Several people can have a phone call speaking their own language, Spoken takes care of translating in their respective languages. Send a message to the spoken.tech chat bot saying "start call with X". Call the given number, start chatting with your colleague / client in your own language, the translation of the second person speaking will appear when he'/she will start talking on your CISCO Spark screen.
How we built it
Leveraging the CISCO Spark and Tropo APIs, we were able to connect two callers with the IBM Watson API in real-time in order to provide a live bi-directional translation service that notified callers via a Spark direct message. Our architecture was mostly based around a series of Node.JS connected webhooks, running on top of Kubernetes inside Docker containers.
Challenges we ran into
Making the translation looks relevant and accurate was tricky.
Accomplishments that we're proud of
- Making the translation accurate
- Building it in less than 24 hours
What we learned
A bit of Spanish
What's next for Spoken.Tech
Integrating more languages
Log in or sign up for Devpost to join the conversation.