This system automates the process of generating professional-grade equity research reports for public companies. It replicates the work of human equity analysts by extracting, parsing, analyzing, and reporting on financial documents such as 10-K, 10-Q, earnings call summaries, valuation metrics, and financial news — all within minutes.
-
Scrapers: Fetch 10-K, 10-Q, earnings call transcripts, market data, and news
-
Analytics Engine: Parse financials, detect risks, calculate key ratios
-
LLM Agents: Multiple agents specialized in financial insight generation, sentiment evaluation, and valuation thesis building
-
Benchmarking Module: Peer comparison and risk deltas across time
-
Report Generator: Automated multi-page analyst brief in Markdown and PDF formats
-
Streamlit UI: Interactive frontend for enterprise users, analysts, and clients
equity-research-agent/
│
├── data/ # Raw + processed financial data
├── scrapers/ # EDGAR, Yahoo Finance, NewsAPI
├── agents/ # LLM-driven analysis agents
├── analytics/ # Parsing logic + ratio calculators
├── report_gen/ # Report rendering (Markdown, PDF)
├── streamlit_ui/ # Frontend dashboard
├── utils/ # Configs, logging, helpers
├── tests/ # Unit tests
├── main.py # Orchestration script
├── README.md # This file
├── requirements.txt # Python dependencies
├── LICENSE # MIT License
- Upload or enter stock ticker
- Choose timeframes or compare filings
- Toggle valuation models (P/E, EV/EBITDA, DCF)
- Download 5–8 page research briefs
- Risk radar dashboard and peer stack comparison
- Python 3.11
- OpenAI API
- SEC EDGAR + Yahoo Finance + NewsAPI
- LangChain
- Pandas, NumPy, Matplotlib
- Streamlit
This project is licensed under the MIT License.
Built and maintained by Krishna Gangwal. Inspired by real-world equity research automation needs.
For enterprise deployment or collaboration: krishnagangwal@icloud.com