SpeakUp

• 

  this is the dashboard

Inspiration

Many students and professionals struggle with public speaking, interviews, group discussions, and communication confidence. Existing platforms either provide only random speaking topics or limited feedback. We wanted to create a personal AI communication coach that helps users practice, receive detailed feedback, and continuously improve their speaking skills in a structured and engaging way.

What it does

SpeakUp AI Coach is an AI-powered communication training platform that helps users improve public speaking and communication skills.

Key features include:

AI-generated speaking topics Real-time speech-to-text transcription AI-powered speech evaluation Confidence, fluency, grammar, and vocabulary scoring Filler word detection Personalized communication coaching Mock interview practice Debate mode with AI-generated counterarguments Progress tracking and analytics dashboard

The platform acts as a personal speaking mentor that provides actionable feedback after every practice session.

How we built it

We built SpeakUp using a modern full-stack architecture:

Next.js 15 for frontend and backend API routes TypeScript for type safety Tailwind CSS and Shadcn UI for a modern user interface Framer Motion for smooth animations Google Gemini API for topic generation, speech evaluation, interview simulation, and communication coaching Web Speech API for real-time speech-to-text transcription Zustand for state management Local Storage for storing user progress and session history Recharts for visualizing performance analytics Challenges we ran into Converting speech into accurate real-time transcripts across different speaking styles and accents. Designing effective Gemini prompts to generate structured and meaningful feedback. Measuring communication quality beyond simple grammar correction. Detecting filler words and calculating meaningful speaking metrics. Creating an intuitive user experience while managing multiple AI-powered features. Optimizing response times for AI-generated evaluations and coaching. Accomplishments that we're proud of Successfully built an AI-powered communication coach instead of a simple topic generator. Implemented real-time speech analysis and feedback. Created personalized AI coaching based on user performance. Developed mock interview and debate practice modes. Designed a professional SaaS-style user experience. Built a scalable architecture that can support future features and integrations.

What we learned

Throughout this project, we learned: How to integrate Large Language Models into real-world applications. Prompt engineering techniques for structured AI outputs. Speech recognition and transcript processing workflows. Designing user-centric AI experiences. State management and scalable frontend architecture. Building analytics-driven feedback systems. The importance of balancing AI automation with actionable user insights. What's next for SpeakUp

Future plans for SpeakUp include:

Voice tone and emotion analysis AI-powered pronunciation assessment Video-based body language evaluation Multi-language speaking practice Live AI interview simulations with voice interaction Personalized learning paths and communication courses Leaderboards and gamification features Cloud database integration for cross-device synchronization Mobile application for Android and iOS Enterprise version for colleges, training institutes, and corporate communication programs

Built With

• analytics
• and
• api
• coaching
• conversion
• lightweight
• local
• management
• next.js
• performance
• progress
• real-time
• recharts
• session
• speech
• speech-evaluation
• speech-to-text
• state
• storage
• typescript
• user
• visualization
• web
• zustand