Born in Argentina and having moved to the US as an 11 year old, language barriers posed a great challenge when trying to make friends, learning in a classroom setting, or while studying audio-visual material. In a world where culture and language are evolving faster than ever before, communication has become increasingly challenging, especially for the visually/audibly impaired and the monolingual. Our goal was to provide a resource which could be repeatedly applied with low engineering costs to mitigate these concerns.
What it does
Our project consists of an open-source API focusing on real-time translations and communication. Through our API, users can interface with multiple Google Cloud Platform (GCP) services to achieve rapid, efficient multilingual communication in their applications. To display the ease of use and integration, we created a Vue website allowing for clear bilingual communication, shattering language barriers in the process. Our website focuses on accessibility for the visually and audibly impaired, providing clear lines of communication with everyone from your great aunt, your new neighbor, or even the new foreign exchange student.
How we built it
We built a node.js backend to work with the GCP SDK and a frontend in Vue to showcase the usability of our API. Utilizing Cloud Build and Cloud Source Repository, we were able to setup CI/CD for our frontend to automatically deploy to App Engine. For our backend, we created a Compute Engine machine to host an nginx and node.js API server that connects to GCP to use their translation and speech APIs.
Challenges we ran into
- Hosting with GCP
- Compute Engine configuration specifics
- Building Services using Terraform
- Layouts with Vue.js and Vuetify
- Choosing a project idea
- Deciding on specific frameworks/languages
- Audio file encoding and FormData HTTP behavior
Accomplishments that we're proud of
- Recording and manipulating audio buffers
- Learning GCP with Terraform
- Connecting GCP services together
- Setting up a CI/CD pipeline
What we learned
- Utilizing GCP services and APIs
- CI/CD Cloud Build and App Engine
- Compute Engine
- GCP IAM Roles and Permissions
- Vue.js, Vuetify, and Figma
- Audio file and buffer manipulation
What's next for Gibbri.sh
Rapidly prototype future applications using our technology to seamlessly connect people to fundamental resources such as emergency hotlines without the fear of miscommunication. Our future goals are to decrease latency for our speech and translation services, and to bring our technology into publicly accessible applications.