VoiceEdge AI

Product Intro
Tech Stack

Inspiration

We were inspired by two key challenges: the widespread anxiety around presentations among students and the unique difficulties faced by people with hearing disabilities in developing effective presentation skills. Recognizing that 75% of students fear presentations and that individuals with hearing impairments struggle to self-correct their speech patterns, we saw an opportunity to create an inclusive tool that empowers both groups to overcome these challenges.

What it does

VoiceEdge AI is an AI-powered speech skills coach tailored for high school and college students, as well as individuals with hearing disabilities. It provides:

Real-time feedback on speech patterns, rhythm, intonation, body language, and content delivery
Personalized coaching to improve storytelling and persuasive skills
Real-time practice sessions with your role model
Visual cues and haptic feedback for users with hearing disabilities
A safe, judgment-free environment for practice and skill development
Adaptive learning that caters to different presentation scenarios and user needs
Mindfulness techniques to help reduce stress and anxiety and build confidence for presentations

How we built it

Test the API on the backend. Local computer we tested out the unit tests and Semantic used the whisper API. Acoustic used the Praat.

Challenges we ran into

We ran into problems with our computers not having the specified libraries, often running into ModuleNotFound errors. When trying to import HumeVoiceClient from the Hume library, we run into this problem. We solved this by using Google Colab instead of our local machines. By doing so, we were able to collaborate with one another on the project and work harmoniously. Debugging these issues meant that we had to spend time looking into forums and documentation. In addition, we also had problems with our access codes. We sought help by going into the Slack channels and using the redeem codes that they provided in the chat. By redeeming the free credits, we were able to focus on building the project and not having to worry about the billing.

Accomplishments that we're proud of

Democratizing Role Models: Leveraged AI to make learning from role models accessible and widespread, reducing the scarcity of real-life mentors. AI for Good: Integrated mindfulness techniques to help users manage stress and anxiety, enhancing their presentation skills. Advanced Feedback: Utilized Hume's speech recognition and large language models (LLMs) like OpenAI and AWS Claude to provide clear, actionable feedback, making it easier for students to improve their skills.

What's next for VoiceEdge AI

Streamline Product (1 Month):
Transition to a fully online platform with real-time feedback capabilities
Optimize system performance to reduce latency for a smoother user experience
User Research & NPO Collaboration (3 Months):
Conduct comprehensive user behavior analysis by recruiting diverse students
Partner with non-profit organizations to engage with hearing-impaired users
Test and refine business models based on feedback from NPO collaborations
Ecosystem & Feature Expansion (6 Months):
Develop an inclusive, engaging ecosystem featuring interactive avatars (e.g., teddy bears) to create a more welcoming practice environment
Introduce new modules for networking skills and small talk practice
Expand the platform's capabilities to cater to a wider range of communication scenarios
Continuous Improvement:
Regularly update our AI models based on user data and feedback
Enhance accessibility features for users with hearing disabilities
Explore partnerships with educational institutions to integrate VoiceEdge AI into curricula
Community Building:
Create a supportive user community for peer feedback and shared experiences
Organize virtual events and workshops to foster engagement and skill development
Research and Development:
Invest in ongoing research to improve AI coaching capabilities
Explore emerging technologies to enhance the user experience and effectiveness of the platform