Professional-grade equity research assistant that turns natural-language questions into clear investment briefings. The system blends multi-agent orchestration, retrieval-augmented generation, and live financial data to produce analyst-style reports in seconds.

• Multi-agent LangGraph workflow coordinates routing, market data, sentiment, visualization, and reporting.
• Real-time analytics pull from Yahoo Finance, SEC EDGAR filings, and curated news.
• RAG pipeline stores vectorized filings and market context in ChromaDB for grounded answers.
• Conversation memory in MongoDB keeps 24-hour session history for follow-up questions.
• Bilingual interface automatically supports English and Chinese queries.

• Backend: FastAPI service with async agents, conversation memory, and ingestion scripts.
• AI Orchestration: LangGraph + LangChain agents backed by OpenAI models.
• Data Services: Yahoo Finance adapters, SEC EDGAR scraper, ChromaDB vector store.
• Frontend: React + TypeScript + Tailwind for charts, reports, and session management.

• Python 3.11+
• Node.js 18+
• MongoDB instance
• OpenAI API key

git clone <your-repo-url> cd multi-agent-investment python -m venv .venv . .venv/Scripts/activate # Windows # source .venv/bin/activate # macOS/Linux pip install -r requirements.txt cd frontend && npm install && cd ..

Copy .env.template to .env and supply:

• OPENAI_API_KEY – OpenAI credential for chat + embeddings
• MONGODB_URI – MongoDB connection string
• MONGODB_DB_NAME – Database name (default investment_research)
• CHROMA_PERSIST_DIR – Vector store path (default ./data/chroma)

python -m backend.scripts.init_db python -m backend.scripts.init_ticker_cache

# Terminal 1 uvicorn backend.main:app --reload --port 8000 # Terminal 2 cd frontend npm run dev

• Web app: http://localhost:3000
• API docs: http://localhost:8000/docs

Web Interface

1. Open the frontend.
2. Ask: “Analyze Apple’s near-term outlook.”
3. Review the generated report, charts, and investor snapshot.

API

curl -X POST http://localhost:8000/api/research/query \ -H "Content-Type: application/json" \ -d '{"query": "What is the investment outlook for NVDA?"}'

backend/ FastAPI app, agents, RAG pipeline, services frontend/ React client with charts and report UI assets/ Product screenshots data/ Local caches (vector store, ticker cache) tests/ Workflow and ingestion checks 

Developer: @Manokil