WORKS
Toby is an innovative AI startup that developed a desktop application for real-time call’s translation and transcription, enabling smooth communication between people who speak different languages. This platform helps global teams communicate and also contributes with companies in finding, hiring, & working with multilingual talent. Its primary objective is to build a world where global teams exist not only beyond physical borders, but beyond language barriers. Toby brings the solution by bridging gaps between multicultural groups of people working together speaking many different languages and making the world feel more united.
Client requirement:
Nearshore.io worked closely with Toby to thoroughly understand their needs and objectives. We analyzed the functional and non-functional requirements, ensuring our solution aligned perfectly with Toby’s vision and technical goals.
We collaborated with their team to bring a demonstration of their application to a web environment, allowing new users to experience a real-time simulation of a call with Toby.
System architecture:
After analyzing various options, Webflow was the platform chosen to implement the project. Despite not being specially designed for web applications, it allows us to integrate all scales without compromising performance and maintaining smooth user experience.
Integration of essential features:
Our development was based on integrating a socket connection between Toby's server and our web application, enabling seamless real-time interaction without delays in responses. We implemented a real-time audio processor to send data from the user's microphone and receive and play the AI’s responses.
The interface we developed is interactive and dynamic, it simulates what the user expects from any video calling application like Meet or Zoom, complete with volume options and visual feedback when the user or AI is speaking.
Website optimization:
Several measures were taken to optimizeToby functionalities. A transcription section was implemented where users are able to read everything said and heard during the call in a stream format similar to GPT. Furthermore, we included a feature that allows users to submit corrections for incorrect transcriptions, helping Toby team further refine their product by leveraging direct user feedback.
This project was a rewarding experience that allowed us to support Toby in showcasing the potential of their innovative product to new users. The close collaboration between Toby’s team and ours led to a high-quality result, effectively demonstrating what their desktop application can offer. This project not only highlighted the power of teamwork but also revealed how thoughtful integration and innovative solutions can bring a visionary product to life in new and engaging ways.
