Instinct provides auto-translated text messages for barrierless communication.
Instinct translates and displays texts in the recipient’s preferred language in real-time, utilizing Twilio and GCP’s Translation API.
Instinct increases speed and coherence of conversations, which is particularly critical during emergencies.
Globalization doesn’t stop, and neither should innovation.
Over the past few years, developments such as Airbnb and more affordable flights have made traveling more accessible than ever. With technology constantly evolving, the world grows physically closer, with countless opportunities for intermingling. As globalization pushes forth, it is clear that our methods for dealing with an interconnected world must also improve.
With this notion in mind, our team brainstormed ways to improve an act seemingly simple and intuitive: communication.
What it does
Instinct is a service that translates text messages in real-time, allowing for a seamless conversation in each user’s native language. Preferred languages are automatically recognized based on each user’s outbound messages, and all incoming messages are automatically translated and displayed in that language.
Features • Easy set-up: Two users enter their phone numbers on a web application and receive an intermediate Twilio number, to be addressed for future (translated) communications between the pair. • Adaptive: Texts will be translated to the language of the sender’s most recent outbound message, thus users can continually change their preferred language. • Offline: Does not require cellular data or internet connection, adding reliability. • Automatic: Translations are accessed and displayed once a text is received, requiring no effort on each user’s part.
Applications • Safety: Instinct was designed with travelers and immigrants in-mind. In emergencies, it is particularly important that non-native speakers are able to relay and receive information as swiftly and coherently as possible. Not only does our service allow users to report more comfortably, its automatic nature also eliminates the need for tedious manual translation apps. Understanding that many tourists and new immigrants may not have access to a roaming plan, we also decided to use native text messaging, which works reliably without internet access. With Instinct, local alerts can additionally be communicated to the public with the trust that it will be understood by all. • Awareness: The first step to awareness is communication. Instinct seamlessly removes language barriers and enables cultural exchange between users. Now, you can have pen pals unimaginably far and different from you, and begin to comprehend them even without a language in common!
How I built it
Our project had the following components: • Flask application that utilizes Twilio APIs to receive and send text messages from an intermediate number. • Google Could Translation APIs to automatically recognize both the sender and recipient’s choice of language as well as convert text. • Lightweight Web app with a supporting SQL database to create and store Twilio intermediate numbers for each conversation.
Challenges I ran into
Automating Twilio services such as buying phone numbers and linking numbers to API endpoints required more creativity and workarounds since we couldn’t use the Twilio web app. We tried to initially set everything up on the cloud platform, however quickly realized that local development was much more suitable for the rapid builds and testing this project required. We also ran into some problems while setting up Instinct’s preset functionality, which allows the service to remember each user’s preferred language and will automatically convert texts to that language without any input by the user. In the beginning, we had some struggles with the SQL database not accepting inputs, but were eventually able to solve it through storing the database as a session variable. Lastly, we also had issues deploying the app through GCP's app engine.
Accomplishments that I'm proud of
Beginning this project, we were uncertain about our prospects. We all came from different colleges, backgrounds, and expertise, and had just met each other at Treehacks. One thing we had in common though, was our lack of experience with web applications and cloud platforms! The whole project was all-in-all a large hands-on learning experience for us, and we are very proud that we came together as a team to present a complete working demo of our initial idea. We explored APIs we weren’t familiar with, reached out for help from various mentors, and created a functioning product easily scalable with our web app and cloud platform.
What I learned
Through our brainstorming and creation process, we learned that the oldest, most basic, most well-integrated technologies in our everyday lives can be improved upon. Picking apart the native text, rather than creating a new, separate application completely, we gave ourselves time to think about opportunities for advancement in-reach. We forced ourselves into the shoes of various users, considering their needs, their pain points, their interests –– a mindset useful in both human-centered innovation and daily life.
There is still much room for improvement in this deceptively simple functionality. To increase accessibility to this service, we could expand to other messaging platforms such as WhatsApp and Messenger. We could additionally enable translations in group chats, as well as in voice messages. On a more challenging note, we could further utilize the power of Twilio APIs to automatically notify users and offer action paths (e.g. calling a certain number) upon detection of suggestive keywords (e.g. “urgent”). These additional safety workflows have much potential for social good, simply by increasing convenience and awareness at critical moments.