URHeard AI

An AI-powered interview platform that transforms hiring through lifelike video interviews. It scales infinitely, eliminates bias, and delivers instant insights - helping companies hire better.

Comment

URHeard AI: Revolutionizing Hiring Through AI-Powered Interviews

What Inspired Us

The inspiration for URHeard AI came from witnessing the broken state of traditional hiring processes. As persons who have been on both sides of the interview table, We noticed several critical pain points:

  • Time Inefficiency: HR teams spending 40+ hours per hire on initial screenings
  • Inconsistent Evaluation: Different interviewers asking different questions, leading to unconscious bias
  • Poor Candidate Experience: Long wait times and impersonal processes driving away top talent
  • Scalability Issues: Companies struggling to handle high-volume hiring while maintaining quality

The name "URHeard" reflects our core mission: ensuring every candidate feels truly heard and evaluated fairly, regardless of who's conducting the interview. We wanted to create a solution that would democratize the hiring process while maintaining the human touch that candidates deserve.

Challenge Compliance

Conversational AI Video Challenge

Used Tavus to bring real-time AI video agents into our Bolt.new app. URHeard AI features lifelike video personas that conduct interviews naturally, enabling a human-like hiring experience at scale.

Startup Challenge

Used Supabase to prepare our Bolt.new project for scale. Supabase powers our authentication and database backend, supporting real-time operations and team-based workflows.

Deploy Challenge

Deployed our full-stack Bolt.new application using Netlify. This ensured continuous deployment, performance optimization, and ease of global accessibility.

What We Learned

Building URHeard AI was an incredible learning journey that taught me several key lessons:

Technical Skills Acquired

  • Real-time Video Integration: Mastered WebRTC technology through Daily.co integration for seamless video interviews
  • AI/ML Integration: Learned to integrate multiple AI services (Tavus API for video personas, Gemini API for analysis)
  • State Management: Implemented complex state management using Jotai for handling interview sessions, user authentication, and real-time data
  • Real-time Communication: Built robust WebSocket connections and message handling for live interview interactions
  • Advanced React Patterns: Implemented custom hooks, context providers, and component composition for scalable architecture

Business & Product Insights

  • User Experience Design: Learned to balance technical complexity with user-friendly interfaces
  • API Integration Strategy: Developed patterns for integrating multiple third-party services while maintaining reliability
  • Data Processing: Built sophisticated transcript analysis and scoring systems
  • Security Considerations: Implemented proper authentication, authorization, and data protection measures

Soft Skills Developed

  • Problem-Solving: Tackled complex technical challenges like real-time transcript capture and AI analysis
  • User-Centric Thinking: Continuously refined the product based on user feedback and needs
  • Technical Architecture: Designed scalable systems that can handle multiple concurrent interviews

How We Built the Project

Technology Stack

  • Frontend: React 18 with TypeScript, Vite for build tooling
  • Styling: Tailwind CSS with custom design system
  • State Management: Jotai for atomic state management
  • Video Integration: Daily.co for WebRTC video calls
  • AI Services: Tavus API for AI personas, Gemini API for transcript analysis
  • Backend: Supabase for database and authentication
  • Payments: Stripe integration for subscription management
  • Animation: Framer Motion for smooth user interactions

Development Process

Phase 1: Core Infrastructure

We started by building the foundational architecture:

  • Set up the React application with TypeScript and modern tooling
  • Implemented authentication system using Supabase
  • Created the basic routing structure and layout components
  • Established the design system with Tailwind CSS

Phase 2: Video Integration

The most complex part was integrating real-time video capabilities:

  • Integrated Daily.co for WebRTC video calls
  • Implemented Tavus API for AI-powered video personas
  • Built real-time transcript capture and processing
  • Created interview session management system

Phase 3: AI Analysis Engine

Developed sophisticated analysis capabilities:

  • Built transcript processing service with multiple AI providers
  • Implemented scoring algorithms for communication, technical skills, and problem-solving
  • Created detailed reporting and insights generation
  • Added export functionality for PDF and other formats

Phase 4: User Experience

Focused on creating an intuitive user experience:

  • Designed responsive dashboard for interview management
  • Built candidate entry and session tracking
  • Implemented real-time progress indicators and notifications
  • Created comprehensive results and analytics views

Phase 5: Business Features

Added essential business functionality:

  • Integrated Stripe for subscription management
  • Built billing and usage tracking systems
  • Implemented team collaboration features
  • Added advanced interview customization options

Key Features Implemented

  1. AI-Powered Video Interviews: Lifelike AI interviewers that conduct natural conversations
  2. Real-time Analysis: Instant scoring and insights during interviews
  3. Bias-Free Evaluation: Objective assessment based on merit and skills
  4. Scalable Architecture: Handle multiple concurrent interviews globally
  5. Comprehensive Reporting: Detailed transcripts, scores, and recommendations
  6. Custom Interview Types: Support for job interviews, discovery calls, and custom scenarios

Challenges We Faced

Technical Challenges

1. Real-time Video Integration Complexity

Challenge: Integrating multiple video services (Daily.co and Tavus) while maintaining seamless user experience Solution: Built abstraction layers and fallback mechanisms, implemented proper error handling and connection management

2. AI Analysis Reliability

Challenge: Ensuring consistent and accurate AI analysis across different interview types and candidate responses Solution: Implemented multiple AI providers with fallback mechanisms, created robust prompt engineering, and built validation systems

3. State Management Complexity

Challenge: Managing complex state across interview sessions, user authentication, and real-time data Solution: Used Jotai for atomic state management, implemented proper state persistence, and created clear data flow patterns

4. Real-time Transcript Processing

Challenge: Capturing and processing transcripts in real-time while maintaining accuracy Solution: Built a sophisticated transcript service with multiple capture methods, implemented proper speaker identification, and created robust error handling

Business Challenges

1. API Integration Complexity

Challenge: Integrating multiple third-party APIs while maintaining reliability and performance Solution: Built comprehensive error handling, implemented retry mechanisms, and created fallback systems

2. User Experience Design

Challenge: Creating an intuitive interface for complex interview management while maintaining simplicity Solution: Conducted user research, implemented progressive disclosure, and created comprehensive onboarding flows

3. Data Security and Privacy

Challenge: Ensuring candidate data security and compliance with privacy regulations Solution: Implemented proper encryption, built secure data handling practices, and created comprehensive privacy controls

Personal Growth Challenges

1. Scope Management

Challenge: Balancing feature development with project scope and timeline Solution: Implemented agile development practices, prioritized core features, and created iterative development cycles

2. Technical Decision Making

Challenge: Making architectural decisions that would scale with the project's growth Solution: Researched best practices, consulted technical documentation, and built flexible, modular systems

Impact and Results

URHeard AI has successfully addressed the core problems that inspired its creation:

  • Reduced Time-to-Hire: Companies can now conduct initial screenings in minutes instead of hours
  • Improved Consistency: Standardized evaluation criteria eliminate interviewer bias
  • Enhanced Candidate Experience: Professional, engaging AI interviews that candidates actually enjoy
  • Scalable Solution: Companies can handle high-volume hiring without sacrificing quality

The platform demonstrates how AI can enhance human processes rather than replace them, creating a more efficient and fair hiring ecosystem.

Future Vision

Looking ahead, we plan to expand URHeard AI with:

  • Advanced AI training capabilities for custom interview personas
  • Integration with popular HR platforms and ATS systems
  • Enhanced analytics and predictive hiring insights
  • Multi-language support for global hiring needs
  • Advanced accessibility features for inclusive hiring

This project has been an incredible journey of learning, problem-solving, and innovation. It's shown us the power of combining cutting-edge technology with human-centered design to solve real-world problems.

Built With

  • ai-powered-transcript-analysis
  • authentication
  • autoprefixer
  • class-variance-authority
  • clsx
  • code-splitting
  • custom-design-system
  • daily.co-api-(video-calls)
  • database)
  • framer-motion
  • frontend-technologies:-react
  • glassmorphism-effects
  • google-gemini-api-(transcript-analysis)
  • jotai
  • json-export
  • lazy-loading
  • live-video-streaming-export-&-reporting:-pdf-generation
  • lucide-react
  • natural-language-processing-payment-&-billing:-stripe
  • postcss
  • postgresql-video-&-communication:-daily.co
  • react-router-dom
  • react-tsparticles
  • real-time-transcript-capture
  • responsive-design-performance-&-optimization:-vite-build-tooling
  • stripe-api-(payments)
  • subscription-management-development-tools:-eslint
  • supabase-api-(database-operations)-state-management:-jotai-(atomic-state-management)-authentication:-supabase-auth-file-storage:-supabase-storage-(inferred-from-database-structure)-real-time-features:-websocket-connections
  • tailwind-css
  • tailwind-css-animate-backend-&-database:-supabase
  • tailwind-merge
  • tavus-api-ai-&-machine-learning:-google-gemini-api
  • transcript-analysis-ui/ux-libraries:-radix-ui
  • typescript
  • typescript-eslint-cloud-services:-supabase-(hosting
  • vercel/netlify-(deployment-inferred)-apis-&-integrations:-tavus-api-(ai-video-personas)
  • vite
  • vite-for-build-tooling-**styling**:-tailwind-css-with-custom-design-system-**state-management**:-jotai-for-atomic-state-management-**video-integration**:-daily.co-for-webrtc-video-calls-**ai-services**:-tavus-api-for-ai-personas
  • webrtc
Share this project:

Updates