Financial Analyzer

Universal Financial Document Intelligence Platform

Transform any financial document into actionable intelligence using advanced LLM technology with RAG (Retrieval-Augmented Generation). Analyze quarterly reports, SEC filings, 10-K/10-Q forms, earnings transcripts, and annual reports from companies worldwide.

Features

Dual-Mode Analysis System

• Single Agent Mode: Fast, unified analysis with structured bullet-point output - ideal for quick insights
• Multi-Agent Mode: Specialized agents (Researcher, Analyst, Risk Assessor, Strategist) for deeper analysis with structured formatting
• Consistent Output: Both modes produce clean, structured bullet points with no markdown formatting
• Intelligent Fallback: Automatic fallback to Single Agent if Multi-Agent encounters issues
• Mode Persistence: Your preferred mode is saved automatically

Universal Document Processing

• Multi-Format Support: Quarterly reports, SEC filings, 10-K/10-Q forms, earnings transcripts
• Global Coverage: Analyze financial documents from companies worldwide
• Smart Chunking: Memory-optimized handling for large regulatory filings

LLM-Powered Analysis

• Advanced RAG: Context-aware information retrieval with vector embeddings
• Structured Output: Automated formatting ensures consistent bullet-point structure across all sections
• Multi-Agent Pipeline: Planner → Researcher → Analyst → Validator → Formatter workflow (Single Agent mode)
• Specialized Agents: Role-based agents with expertise-specific analysis (Multi-Agent mode)
• Parallel Processing: Simultaneous analysis of multiple report sections
• Error Resilience: Automatic retry logic with intelligent rate limiting
• Strategy Pattern: Modular architecture supporting multiple analysis approaches

Generated Analysis Sections

All sections are generated with structured bullet points (•) for easy readability:

• Company Overview: Business model, operations, market position, and strategic initiatives
• Financial Highlights: Revenue trends, profitability metrics, ratios, and performance indicators with specific numbers
• Risk Assessment: Business, operational, financial, regulatory, and market risks with impact analysis
• Management Commentary: Executive outlook, strategic priorities, forward guidance, and long-term vision

Interactive Q&A

• Intelligent Chat: Ask questions about the analyzed document
• Context-Aware Responses: Answers based on actual document content
• Semantic Search: Advanced retrieval for precise information

Technology Stack

Frontend

• React 18.3.1 with TypeScript 5.5.3
• Vite 5.4.1 - Fast build tool and dev server
• Tailwind CSS - Utility-first CSS framework
• shadcn/ui - Modern UI component library

Backend

• Node.js/Express - REST API backend
• OpenRouter - Access to multiple LLM providers including GPT-4 Turbo
• HuggingFace - Embeddings for semantic search
• Vector Store - In-memory semantic search with HuggingFace embeddings

Running the Application

1. Install dependencies for both frontend and backend:

# Install backend dependencies
cd backend
npm install

# Install frontend dependencies
cd ../frontend
npm install

2. Start the backend server:

cd backend
npm run dev

The server will start on http://localhost:3001

3. Start the frontend development server:

cd frontend
npm run dev

Access the application at http://localhost:5173

Environment Variables

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

# Required for OpenRouter (LLM API access)
OPENROUTER_API_KEY=your_openrouter_api_key_here

# Optional: HuggingFace API key for faster embeddings (fallback to local model if not provided)
HF_API_KEY=your_huggingface_api_key_here

Getting API Keys

1. OpenRouter API Key:

   • Sign up at OpenRouter
   • Get your API key from the keys page
   • Add funds to your account (required for API usage)

2. HuggingFace API Key (Optional):

   • Sign up at HuggingFace
   • Get your API key from settings
   • The app will work without this, but with slower local embeddings

Project Structure

Frontend

src/
├── components/           # React components
│   ├── ui/              # Reusable UI components
│   ├── ChatInterface.tsx
│   ├── FileUpload.tsx
│   ├── ReportGenerator.tsx
│   └── ReportDisplay.tsx
├── hooks/               # Custom React hooks
├── pages/               # Application pages
└── lib/                 # Utility libraries

Backend

backend/
├── server.js                      # Express server & API endpoints
├── ragOrchestrator.js             # Strategy pattern orchestrator
├── agenticWorkflow.js             # Agentic pipeline with structured formatting
├── strategies/
│   ├── singleAgentStrategy.js    # Single agent with planner-researcher-analyst-validator-formatter pipeline
│   └── multiAgentStrategy.js     # Multi-agent with specialized roles and structured output
├── pdfProcessor.js                # PDF text extraction
├── package.json                   # Dependencies
└── uploads/                       # Temporary file storage

API Endpoints

Document Processing

• POST /api/upload - Upload financial document (PDF)
• POST /api/generate-report - Generate comprehensive financial analysis (supports mode parameter: single or multi)
• GET /api/fetch-pdf - Fetch PDF documents from external URLs

Interactive Analysis

• POST /api/ask-question - Ask questions about analyzed documents (uses same mode as report generation)
• GET /api/rag-modes - Get available analysis modes and their status
• GET /api/health - Check API health and environment status

Supported Documents

• Quarterly Reports (10-Q, Q1/Q2/Q3/Q4)
• Annual Reports (10-K, Annual Filings)
• SEC Filings (8-K, Proxy Statements)
• Earnings Materials (Call Transcripts, Presentations)
• Credit Reports (Rating Agency Reports)

---

Financial Analyzer - Transform financial documents into actionable intelligence with LLM precision.