RightHand

RightHand Logo

Inspiration

As students, we were inspired by the low amount of up to date information on university courses and lectures. So we wanted to create something that could collect more of this data, while making it seamless, natural, and similar to having a typical conversation with a fellow peer.

What it does

RightHand is an AI-powered voice assistant that replaces boring course evaluation forms with a real, natural conversation. Instead of forcing students to fill out tedious bubbles and text boxes, RightHand acts as a digital peer that "catches up" with you after a lecture.

By simply speaking, students can share their thoughts on the lecture's clarity, the professor's engagement, or the difficulty of the material. RightHand then uses Natural Language Processing (NLP) to extract key insights from the conversation such as sentiment, specific pain points, and highlights to automatically generate a comprehensive, course experience summary. It turns casual venting or praise into valuable, structured data for the student community.

How we built it

To bring RightHand to life, we decided to integrate the OpenAI API to power the "brain" of our assistant, enabling it to understand student feedback and extract data from messy, natural dialogue. For the voice itself, we integrated the ElevenLabs API to provide a realistic, human-like voice that makes the interaction feel more like a genuine conversation rather than a robotic interview. The entire experience is housed within a custom-built stack using JavaScript, managing both the interactive frontend and the backend logic, styled with CSS to ensure a clean, intuitive student interface.

Challenges we ran into

The biggest challenge was fine-tuning the AI’s conversational flow. Initially, the assistant struggled with any unpredictability in human speech. For example, if a student gave a vague answer or went off topic, the AI would often give an unnatural response or awkwardly jump to the next scripted question. We had to work a lot on the prompt engineering to ensure the AI could "pivot" smoothly and acknowledge the student's words before guiding the conversation back to the data it needed to collect.

Accomplishments that we're proud of

As most of us are first-time hackathon participants, our proudest achievement was successfully taking RightHand from a concept on a whiteboard to a fully functional prototype. Despite the steep learning curve and the technical frustrations, we pushed through to create a tool that actually works, and in the end it made all the hours of troubleshooting worth it.

What we learned

In the end, this weekend was almost like a quick crash course on rapid software development. We learned what it takes to build a product from the ground up in just 48 hours, from leveraging different APIs to debugging frontend glitches under pressure. Above all, we experienced the crazy environment of a hackathon and learned how to stay focused even when things got frustrating.