JustifyAI

Intelligent Fact-Checking and Content Enhancement Platform

---

---

Transform AI responses into verified, trustworthy content with intelligent fact-checking and visual enhancement

JustifyAI is a comprehensive platform that combines LLM auditing with intelligent content enhancement. It systematically verifies AI-generated responses against real-world information and transforms them into visually appealing, trustworthy content.

• AI models generate inaccurate or misleading information
• Users lack systematic verification of AI-generated content
• Raw AI responses lack visual appeal and proper structure
• No reliable method to validate AI claims against factual data

JustifyAI provides an automated fact-checking layer that:

• Analyzes AI responses for verifiable claims
• Verifies information using web search and knowledge bases
• Generates detailed audit reports with accuracy assessments
• Enhances content with visual formatting and images
• Rewrites responses to correct inaccuracies

• Multi-Agent Architecture: Specialized critic and reviser agents
• Web Search Integration: Real-time verification using Google Search
• Claim Extraction: Automatic identification of verifiable statements
• Accuracy Scoring: Detailed accuracy assessments and reporting

• Visual Formatting: Structured, engaging content presentation
• Image Generation: Relevant visuals to accompany responses
• Professional UI: Clean, modern interface with smooth animations
• Responsive Design: Seamless experience across all devices

• RESTful API: Easy integration with existing applications
• Cloud-Native: Built on Google Cloud Platform for scalability
• Real-time Processing: Fast response times with streaming capabilities
• Comprehensive Logging: Detailed monitoring and debugging support

• Next.js 15.2.4 - React framework with App Router
• TypeScript - Type-safe development
• Tailwind CSS - Utility-first styling
• Framer Motion - Smooth animations
• Radix UI - Accessible component library
• Firebase - Authentication and real-time features

• FastAPI - High-performance Python web framework
• Google Vertex AI - Advanced AI and ML capabilities
• Google Cloud Run - Serverless container platform
• Pydantic - Data validation and serialization
• Uvicorn - ASGI server for production

• Google ADK - Agent Development Kit
• Google Gemini - Large Language Model
• Multi-Agent System - Specialized fact-checking agents
• Web Search Tools - Real-time information verification

• Node.js 18+ and npm/pnpm
• Python 3.10+
• Google Cloud Platform account
• Google Cloud CLI installed

1. Clone the Repository

git clone https://github.com/your-username/genexhange1.git cd genexhange1

2. Backend Setup

cd backend/llm-auditor poetry install # Set up environment variables export GOOGLE_GENAI_USE_VERTEXAI=true export GOOGLE_CLOUD_PROJECT=your-project-id export GOOGLE_CLOUD_LOCATION=us-central1 # Authenticate with Google Cloud gcloud auth application-default login

3. Frontend Setup

cd frontend npm install npm run dev

4. Run the Application

# Backend (Terminal 1) cd backend/llm-auditor poetry run python main.py # Frontend (Terminal 2) cd frontend npm run dev

Visit http://localhost:3000 to access the application.

graph TD A["User Input"] --> B["LLM Auditor API"] B --> C["Critic Agent"] C --> D["Web Search Verification"] D --> E["Accuracy Assessment"] E --> F["Reviser Agent"] F --> G["Content Enhancement"] G --> H["JustifyAI Processing"] H --> I["Visual Enhancement"] I --> J["Final Response"] style A fill:#e1f5fe style J fill:#c8e6c9 style C fill:#fff3e0 style F fill:#fff3e0 

1. Input Processing: User submits query or statement for verification
2. Claim Extraction: Critic agent identifies verifiable claims
3. Web Verification: System searches for supporting evidence
4. Accuracy Assessment: Claims evaluated and scored for accuracy
5. Content Revision: Reviser agent corrects inaccuracies
6. Visual Enhancement: JustifyAI adds formatting and structure
7. Final Output: User receives verified, enhanced content

• Verify historical facts and scientific claims
• Enhance learning materials with visual aids
• Create trustworthy study resources

• Fact-check articles and reports
• Generate verified news summaries
• Create engaging visual content

• Verify product information and claims
• Enhance marketing content with credibility
• Generate trustworthy customer communications

• Validate research findings
• Enhance technical documentation
• Create reliable reference materials

"Earth is further away from the Sun than Mars." 

FACT-CHECKED RESPONSE INACCURATE CLAIM DETECTED Original: "Earth is further away from the Sun than Mars." CORRECTED VERSION "Mars is further away from the Sun than Earth." VERIFICATION DETAILS • Earth's average distance: ~150 million km • Mars' average distance: ~228 million km • Accuracy Score: 0% (Original) → 100% (Corrected) SOURCES VERIFIED • NASA Solar System Data • Astronomical Almanac • Multiple scientific databases 

POST /query Content-Type: application/json { "message": "Your text to verify", "user_id": "optional_user_id", "generate_image": true }

{ "response": "Verified and enhanced content", "session_id": "unique_session_id", "justifyai_response": { "formatted_content": "Enhanced visual content", "images": ["generated_image_urls"], "accuracy_score": 95, "verification_details": {...} } }

# Deploy backend cd backend/llm-auditor poetry install --with deployment python3 deployment/deploy.py --create # Deploy frontend cd frontend npm run build # Deploy to your preferred platform

# Build and run with Docker docker build -t justifyai-backend ./backend/llm-auditor docker run -p 8080:8080 justifyai-backend

We welcome contributions through:

1. Bug Reports: Report issues and bugs
2. Feature Requests: Suggest new features and improvements
3. Code Contributions: Submit pull requests for enhancements
4. Documentation: Help improve documentation and examples

# Fork the repository git clone https://github.com/your-username/genexhange1.git cd genexhange1 # Install development dependencies cd backend/llm-auditor poetry install --with dev cd frontend npm install # Run tests cd backend/llm-auditor python3 -m pytest tests

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

• Google Cloud Platform for infrastructure support
• Google ADK for the agent development framework
• Next.js Team for the React framework
• FastAPI for the Python web framework
• All Contributors who help improve JustifyAI