HireVox

HireVox – Your AI-Powered Recruiter for Smarter, Faster, and Fairer Hiring

---

About HireVox

HireVox is an AI-powered recruitment platform that streamlines the hiring journey — from resume parsing to AI-driven interviews and feedback generation. Recruiters simply provide a job description, and HireVox handles the rest: generating smart questions, scheduling interviews, conducting voice-based assessments, scoring candidates fairly, and preventing cheating through webcam anomaly detection and tab-switch monitoring. With real-time analytics, and voice verification, HireVox empowers companies to hire faster, smarter, and without bias.

This project was developed by Team CodeBlooded and submitted for HackOdisha + 2025.

Live here – Live link
Project Video Summary – Watch here
Slides (PPT) - PPT link

Deployment on Akash

Akash Wallet Address:

akash18czr8t9lmdvn8vwuzqxz63evy3s0z5t32gjrnc

• SDL File: See deploy.yml in the repo.

• The project is deployed and hosted using Akash Network for decentralized GPU cloud compute.

Features

• Resume Parsing & Ranking with AI
  Automatically extracts and ranks candidate resumes to save recruiters hours of manual work.

• AI-Powered Interview (Voice)
  Conducts structured interviews in real-time using AI voice agents, ensuring consistent and unbiased questioning.

• Voice Verification
  Matches candidate’s voice with pre-recorded samples to verify identity and prevent impersonation during interviews.

• Smart Scoring & Feedback
  Analyzes candidate responses, generates unbiased scores, and provides recruiters with structured feedback and recommendations.

• Recruiter Dashboard with Analytics
  Centralized dashboard to view candidate rankings, insights, and hiring metrics.

• Cheat Prevention & Anomaly Detection

• Warns candidates if they switch tabs twice; ends interview on the third attempt.

• Uses webcam anomaly detection to flag unusual activities (e.g., multiple faces, frequent distractions, mobile usage).

---

Tech Stack

• Frontend: Next.js + TailwindCSS
• Backend: Next.js + gofr (for resume parsing), Flask (Python) for voice verification APIs
• AI Models:
  • Akash ChatAPI (Meta-Llama-3-1-8B-Instruct-FP8) → Question generation
  • Vapi Voice Agent → AI-powered voice interviews
  • OpenRouter API (GPT-3.5-Turbo) → Feedback & scoring
• Other Services:
  • Supabase → Database & authentication
  • face-api.js → Webcam anomaly detection
• Cloud: Akash Networks

---

Workflow / How It Works

Recruiter Side

1. Enters job title, description, interview duration, and question type.
2. AI generates interview questions (editable by recruiter).
3. Recruiter shares interview link via email/dashboard.
4. Receives candidate reports with AI-generated scores, feedback & recommendations.

Candidate Side

1. Opens interview link → fills name, email, and uploads resume.
2. AI generates role + resume-specific questions.
3. Candidate joins full-screen AI interview (voice/chat).
4. System verifies candidate’s voice to prevent impersonation.
5. Responses analyzed → feedback & scoring generated.
6. Cheating prevention system monitors tab-switching and webcam anomalies.

---

Getting Started Locally

1️. Clone the Repository

git clone https://github.com/your-uername/hack-odisha-recruit.git

2️. Install Frontend Dependencies

cd HireVox
npm install

3️. Run the Frontend

npm run dev

4️. Install Backend Dependencies

cd resume-parser
npm install

5️. Run the Backend (gofr)

go run main.go

6️. Setup & Run Python Backend (Voice Monitor)

cd ./voice-monitor

Create a virtual environment:

python -m venv .venv

Activate the virtual environment:

• Windows (PowerShell)

  .\.venv\Scripts\Activate

• Linux / macOS

  source .venv/bin/activate

Install dependencies:

pip install -r requirements.txt

Run the Python voice server:

python voice_server.py

Environment Variables

Create a .env.local file in the frontend directory with the following variables:

NEXT_PUBLIC_SUPABASE_URL=<your-supabase-url>
NEXT_PUBLIC_SUPABASE_ANON_KEY=<your-supabase-anon-key>

OPENROUTER_API_KEY=<your-openrouter-api-key>

NEXT_PUBLIC_HOST_URL=http://localhost:3000/interview

NEXT_PUBLIC_VAPI_API_KEY=<your-vapi-api-key>

AKASH_API_KEY=<your-akash-api-key>

For Akash Deployment Configuration (set these in your environment):

# Akash Deployment Configuration
AKASH_KEY_NAME=your_key_name
AKASH_NODE=https://rpc.akash.network:443
AKASH_CHAIN_ID=akashnet-2
AKASH_ACCOUNT_ADDRESS=your_akash_wallet_address

# Docker Configuration
DOCKER_USERNAME=<docker-username>
DOCKER_PASSWORD=<docker-password>

Future Enhancements

• AI-based video interviews with facial expression analysis
• Real-time recruiter notifications
• Mobile app for candidate interviews
• Advanced analytics & predictive hiring metrics
• Marketplace integrations with Workday, Lever, and Greenhouse

---

Team [CodeBlooded]

• [Monalisha Patra] – developer
• [Puneet kumar dhal] – developer
• [Utkal Kumar Das] – developer
• [RudraMadhab Das] – developer

---

Contributing

We welcome contributions! Please fork this repo, create a feature branch, and open a pull request.

---

License

This project is licensed under the MIT License.

---

Acknowledgements

• Akash for ChatAPI (LLM integration)
• Vapi for AI voice agents
• OpenRouter for GPT-3.5 feedback generation
• Supabase for database & auth
• Akash Networks for decentralized cloud hosting
• face-api.js for webcam anomaly detection

---

Hire smarter, not harder — with HireVox!