Technical Details

The development of involved a range of cutting-edge technologies. We chose NodeJS as the backend technology due to its lightweight nature, ease of use, and scalability, allowing us to create a RESTful API that integrates with OpenAI's assistant API to process natural language inputs. To enable fast and seamless responses, we utilized Elevenlabs' streaming API, which allows us to stream audio directly to the frontend. This approach enables users to receive instant voice responses, making the overall experience more engaging and interactive.

In addition, our code was designed to generate a unique user ID on the fly, allowing us to store the full conversation as reference on the backend. This enables the chatbot to provide context-sensitive responses and maintain a comprehensive understanding of the conversation throughout. The simplicity of our implementation is further highlighted by the fact that only a single script tag is required to load the chatbot, making it easy to install on any website.

Testing

We are currently being testing on a UCLA professor's website, providing an innovative way for students, colleagues, and fans to engage with the professor's work. The chatbot has received positive feedback from users, who appreciate its ability to answer questions about research papers and provide insights into the professor's career.