CandidateAI - Local AI-Powered Candidate Evaluation

100% Free. 100% Private. 100% Local.

Evaluate candidates using local AI models - no API tokens, no subscriptions, no data leaving your machine.

🚀 Quick Start (5 Minutes)

1. Install Ollama

Download from ollama.ai or:

# Windows: Download installer from website
# Mac/Linux:
curl -fsSL https://ollama.ai/install.sh | sh

2. Download AI Model

ollama pull qwen2.5-coder:14b

3. Start Backend

# Windows
start.bat

# Or manually:
python api_server.py

4. Start Frontend

cd gitverified-web
npm install
npm run dev

5. Open http://localhost:3000/engine

🏗️ Architecture

Frontend (Next.js :3000)
    │
    ▼
Backend (Python :3001)
    │
    ├── Integrity Agent
    ├── Code Quality Agent
    ├── Uniqueness Agent
    └── Relevance Agent
         │
         ▼
    Ollama (qwen2:1.5b :11434)

📊 What Gets Evaluated

Agent	Score	What It Checks
Integrity	0-10	Resume authenticity, hidden text, keyword stuffing
Code Quality	0-100	Security, best practices, documentation
Uniqueness	0-10	Original work vs tutorial clones
Relevance	0-10	Job requirements match

💰 Cost Comparison

Solution	Cost	Privacy
Traditional SaaS	$0.10+/eval	❌ Data shared
CandidateAI	$0 forever	✅ 100% local

🔧 Optional: Kestra Workflows

For parallel processing of multiple candidates:

docker-compose -f docker-compose.hybrid.yml up -d
# Access Kestra UI: http://localhost:8080

📁 Project Structure

candidateai/
├── api_server.py         # Python API server
├── agents/               # AI evaluation agents
│   ├── hybrid_model.py   # Ollama integration
│   ├── integrity.py
│   ├── code_quality.py
│   ├── uniqueness.py
│   └── relevance.py
├── gitverified-web/      # Next.js frontend
├── flows/                # Kestra workflows
├── start.bat             # Windows quick start
└── docker-compose.hybrid.yml

📄 License

MIT License - Use freely in your hiring process.

---

Built for engineers who value privacy and hate subscriptions.