SpeakUp

A digital space where voices find confidence.

Inspiration

I’ve always noticed how many people — including myself — have thoughts worth sharing but struggle to express them clearly. Whether it’s public speaking, interviews, or even day-to-day conversations, the fear of speaking up holds people back.

The idea for SpeakUp came from that realization — and from my own frustration of not finding a safe, non-judgmental space to practice speaking. I wanted to build something that helps people articulate their ideas with confidence and clarity.

What it does

SpeakUp is a web platform that empowers people to improve their English communication and confidence through AI-driven speech practice, feedback, and daily speaking prompts. It transforms language learning into an interactive, real-world experience — going beyond grammar to focus on articulation, fluency, and confidence, which are crucial in education, professional growth, and global communication.

Users can sign up and access a personalized dashboard.
First-time users are given an initial speaking assessment to gauge their comfort level.
Based on that, the app provides custom prompts — from casual topics to persuasive speeches.
Instead of live sessions, users upload voice recordings, ensuring better audio quality and flexibility.
The platform then provides AI-powered feedback on clarity, tone, pacing, filler words, and overall confidence.

How AI is applied to education

SpeakUp integrates multiple AI components to create a truly intelligent learning experience:

Speech-to-text & speech analysis (Whisper): Evaluate fluency, pronunciation, and pacing.
Sentiment & tone analysis models: Detect confidence and emotional delivery.
LLM-based feedback (Gemini) : Provide natural, conversational feedback that feels like a real coach.
Personalization logic: Uses user data and past results to recommend prompts that match skill levels.

Together, these AI systems transform speaking practice into a dynamic, data-driven learning journey, redefining how communication skills are taught.

How I built it

Frontend: Built using React.js and Typescript for a smooth, modern UI and seamless navigation.
AI Feedback System: Uses speech analysis APIs (like Whisper/OpenAI + sentiment/tone analysis models) to evaluate user recordings and Google Gemini for feedback
Database: Supabase for user profiles, progress tracking, and prompt history.
Design: Minimal, empowering aesthetic — clean colors, rounded edges, and gentle animations to make users feel calm and supported.

Challenges I ran into

Integrating speech analysis models was tricky, especially aligning AI feedback to human-level nuance.
Ensuring privacy for user recordings — voice data is sensitive, so encryption and temporary storage were important.
Structuring the personalization logic (deciding which prompts suit which skill level) took thoughtful tuning.
UI/UX balancing act — we wanted the app to feel friendly, not intimidating.

Accomplishments that I'm proud of

Creating a judgment-free space where users can express themselves freely.
Getting the AI feedback system to actually give insightful comments that feel human.
Designing a product that aligns deeply with SDG 1: No Poverty, by helping people unlock opportunities through communication and self-confidence.

What I learned

Building something meaningful means listening — to users, to your own frustrations, and to what’s missing in the world.
Combining AI with empathy creates powerful results — technology doesn’t just automate, it can empower.
Simplicity in design can be far more impactful than complexity.

What's next for SpeakUp

Add community challenges — users can participate in themed speaking weeks or debates.
Introduce AI voice coaches that give real-time tips during practice.
Launch a mobile app version for accessibility.
Partner with schools and youth programs to bring public speaking training to everyone.