Resonance AI

Resonance AI is designed to help individuals refine their speech delivery through real-time AI-powered feedback. Whether you’re practicing for a presentation, interview, or public speaking event, Resonance AI provides detailed insights on clarity, tone, pacing, filler words, and engagement.

Inspiration Public speaking is a crucial skill, but many struggle with confidence, pacing, and excessive filler words like "um" and "uh". Inspired by the need for instant, AI-driven feedback, we built Resonance AI to assist speakers in real-time speech analysis and improvement.

How I Built It Frontend: Framework: Next.js & TailwindCSS– for server-side rendering and efficient UI updates Speech Recognition: Web Speech API – for real-time speech-to-text conversion State Management: React Hooks (useState, useEffect) – for handling UI states Backend: Developed with Node.js and Express for efficient API handling. Clerk: User authentication and sign in AI Model: Integrated Gemini 1.5 Flash for advanced speech analysis. Speech Recognition: Utilized Web Speech API for real-time transcription. Database: Stored user data securely using Drizzle ORM Speech Transcription Display: Shows the real-time transcribed speech in a UI section.

Challenges & Learnings One of the main challenges was ensuring accurate transcription and meaningful feedback. By integrating Gemini AI, we fine-tuned the model to provide structured, actionable insights rather than generic responses.

Future Improvements We plan to enhance emotion detection, audience engagement analysis, and multilingual support to make Resonance AI even more powerful.

Built With

• clerk
• gemini
• javascript
• nextjs
• node.js
• react
• tailwindcss