Inspiration
Stop Stuttering was inspired by the everyday struggle many people face when speaking publicly, explaining ideas, or communicating under pressure. Our team noticed that while there are many tools for coding practice, very few tools help beginners improve their communication clarity. We wanted to build something simple, accessible, and beginner-friendly that gives users instant feedback on how they sound.
What it does
How we built it
Using Python, JavaScript, and the OpenRouter API, we built a system that records a 30-second audio clip and analyzes the transcript for filler words, clarity, and coherence. We designed a custom scoring algorithm that highlights communication strengths and weaknesses in real time. Building our own filler-word detector taught us a lot about natural-language patterns, regex, and audio processing.
Challenges we ran into
We also learned how challenging it can be to integrate audio input with an external AI model, especially with timing, formatting, and API request limits. One of our biggest challenges was structuring the backend so it could evaluate spoken answers consistently across different prompts. Another challenge was coordinating frontend, backend, and scoring logic across multiple teammates with different experience levels.
Accomplishments that we're proud of
Through this project, we gained hands-on experience with teamwork, version control, API calls, and data-driven evaluation. Most importantly, we learned how to turn a personal pain point into a working tool that helps people communicate more confidently. Stop Stuttering represents the first step toward an AI-powered communication coach that anyone can use.
Log in or sign up for Devpost to join the conversation.