CapitalForge AI

Multi-agent finance research team by Sharad Khare.

CapitalForge AI combines a Web Agent and a Finance Agent into a coordinated Agno team — delivering live market data, analyst ratings, company news, and web research through an AgentOS chat interface with SQLite conversation history.

Built and maintained by Sharad Khare — AI strategist, full-stack developer, and creator of practical agent workflows.

---

What this project does

CapitalForge routes finance questions to the right specialist agent:

1. Web Agent — searches the web via DuckDuckGo for market news, macro trends, and company context

2. Finance Agent — pulls live stock data, analyst recommendations, company info, and news via YFinance

3. Team Coordinator — delegates tasks across both agents and synthesizes unified responses

All conversations are persisted in SQLite for multi-turn research sessions.

---

Why Sharad Khare built CapitalForge AI

Financial research spans two domains — live market data and broader web context. CapitalForge demonstrates how an Agno agent team with specialized tools outperforms a single general-purpose agent for finance workflows.

---

How it works

User Question
   │
   ▼
CapitalForge Team (OpenAI GPT-4o)
   ├── Web Agent → DuckDuckGo search
   └── Finance Agent → YFinance (price, analysts, news, company info)
   │
   ▼
SQLite History (agents.db)
   │
   ▼
AgentOS Web UI

Agent capabilities

| Agent | Tools | Use cases |

|-------|-------|-----------|

| Web Agent | DuckDuckGo | Market news, macro trends, sector research |

| Finance Agent | YFinance | Stock prices, analyst ratings, company fundamentals, news |

YFinance tools included

• Current stock price
• Analyst recommendations
• Company info
• Company news

---

Features

• Dual-agent team: web search + financial data
• Agno Team coordination with GPT-4o
• YFinance live market data and analyst ratings
• DuckDuckGo web search for broader context
• SQLite conversation history across sessions
• AgentOS web interface for conversational research
• Markdown tables for structured financial output

---

Quick start

cd capitalforge-ai
pip install -r requirements.txt
export OPENAI_API_KEY=your_key_here
python app.py

Open the AgentOS URL shown in the terminal (typically http://localhost:7777).

Required API key

• OpenAI API key — set OPENAI_API_KEY as an environment variable

---

Example prompts

• "What is NVDA's current price and latest analyst recommendations?"
• "Compare AAPL and MSFT — fundamentals, news, and analyst ratings"
• "Search for recent news about the semiconductor sector and summarize key stocks"
• "Get Tesla's company info and find recent web articles about its earnings"

---

Project structure

capitalforge-ai/
├── app.py
├── capitalforge/
│   ├── config.py
│   └── team.py
├── requirements.txt
├── pyproject.toml
└── README.md

---

Who this is for

• Investors combining market data with web research in one workflow
• Developers learning Agno multi-agent team patterns
• Finance analysts prototyping AI research assistants
• Anyone wanting a self-hosted web + finance agent team

Use cases

• Pull live stock prices alongside recent web news
• Compare multiple tickers with analyst ratings and fundamentals
• Research sector trends with web search plus financial data
• Maintain multi-turn finance research sessions with SQLite history

---

Disclaimer

CapitalForge AI provides research for informational purposes only. It is not financial advice. Always do your own due diligence before making investment decisions.

---

Author

• Website: sharadkhare.in

License

MIT © Sharad Khare