Financial Agent for Indian Markets

A hands-on project showing how to build an AI-powered financial analysis agent for Indian equities using AngelOne's SmartAPI and Claude (Anthropic's API).

You feed it a stock ticker. It pulls live market data from NSE/BSE, runs it through Claude, and gives you a structured analysis — trend direction, support/resistance levels, volume analysis, and a plain English summary. There's also an enhanced version that cross-references options chain data (IV, OI, Greeks) against price action.

This is not a trading bot. It's a starting point to show what's possible when you combine a free brokerage API with an LLM. Think of it as building blocks for something bigger.

---

Why AngelOne SmartAPI?

Most Indian market data APIs either charge a monthly fee or lock useful features behind paid tiers. AngelOne's SmartAPI is genuinely free — you just need a trading account (which you probably already have if you trade Indian markets).

What you get for free:

• Live NSE/BSE quotes — LTP, OHLC, bid/ask
• Historical candles — 1min to daily, going back years
• Full option chains — with Greeks (delta, gamma, theta, vega) and IV
• WebSocket streaming — real-time tick data
• Order placement — buy, sell, modify, cancel via API
• Portfolio & positions — current holdings and P&L

No rate limit gotchas for normal usage. No surprise bills. It just works.

---

What's in this repo

File	What it does
connect.py	Authenticates with SmartAPI, gives you a session
fetch_data.py	Pulls 30 days of daily OHLCV candles for any stock
agent.py	The main agent — feeds price data to Claude, gets structured analysis
agent_with_options.py	Enhanced agent that also pulls option chain data for deeper analysis
place_order.py	Reference script for placing orders (dry-run by default, won't execute)
config.py	Loads credentials from .env

---

Prerequisites

• Python 3.8+
• An AngelOne trading account
• An Anthropic API key (for Claude)

---

Setup

1. Get your AngelOne SmartAPI credentials

1. Go to smartapi.angelbroking.com
2. Sign up / log in with your AngelOne client ID (the one you use to trade)
3. Go to My Apps → Create an App
   • App name: anything you want (e.g., "financial-agent")
   • Redirect URL: https://localhost (doesn't matter for API usage)
4. You'll get an API Key — save it
5. Enable TOTP on your AngelOne account:
   • Open the AngelOne mobile app → Settings → Security → Enable TOTP
   • You'll get a TOTP secret (a long base32 string) — save this, you'll need it for automated login
   • If you already use Google Authenticator or similar, the secret is what you scanned as a QR code. You may need to re-enable TOTP to see the raw secret.

You'll need these four things:

• API Key (from SmartAPI dashboard)
• Client ID (your trading login ID, like A12345678)
• Password (your trading PIN)
• TOTP Secret (the base32 string, NOT the 6-digit code)

2. Get your Anthropic API key

1. Go to console.anthropic.com
2. Sign up or log in
3. Go to API Keys → Create Key
4. Save the key

3. Clone and install

git clone https://github.com/AM1403x/financial-agent-india.git
cd financial-agent-india

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

pip install -r requirements.txt

4. Configure credentials

cp .env.example .env

Open .env and fill in your actual values:

ANGELONE_API_KEY=your_smartapi_key_here
ANGELONE_CLIENT_ID=A12345678
ANGELONE_PASSWORD=1234
ANGELONE_TOTP_SECRET=your_totp_base32_secret
ANTHROPIC_API_KEY=sk-ant-...

5. Test the connection

python connect.py

If everything is set up correctly:

Generating TOTP...
Authenticating with AngelOne SmartAPI...
Login successful!
Session token: eyJhbGci...
Feed token: abc123...
Connected as client: A12345678

If it fails, you'll get a clear error message telling you what went wrong.

---

Usage

Fetch stock data

# Default: RELIANCE
python fetch_data.py

# Specify a stock
python fetch_data.py --symbol INFY --token 1594

Output:

Fetching 30 days of daily data for RELIANCE...

Date        Open      High      Low       Close     Volume
──────────────────────────────────────────────────────────────
2025-01-15  1285.00   1298.50   1278.30   1292.45   8234521
2025-01-16  1293.00   1305.75   1290.10   1301.20   9123456
...

Run the analysis agent

# Default: RELIANCE
python agent.py

# Analyze a different stock
python agent.py --symbol TCS --token 11536

Output:

Connecting to AngelOne SmartAPI...
Fetching 30 days of data for RELIANCE (token: 2885)...
Sending data to Claude for analysis...

══════════════════════════════════════════════
  ANALYSIS REPORT: RELIANCE
══════════════════════════════════════════════

  Trend:          BULLISH
  Confidence:     High

  Support:        ₹1,278.30
  Resistance:     ₹1,342.80

  Avg Volume:     9,234,567
  Volume Trend:   Increasing

  Summary:
  RELIANCE has been in a steady uptrend over the past 30 days,
  gaining ~5.2%. The stock has formed higher lows consistently
  with strong volume support. Key support at ₹1,278 has held
  on multiple tests. Watch for breakout above ₹1,342 resistance.

══════════════════════════════════════════════

Run the enhanced agent (with options data)

python agent_with_options.py --symbol RELIANCE --token 2885

This gives you everything above plus options market analysis — what IV and OI patterns suggest about market expectations.

---

How to find symbol tokens

SmartAPI needs a "symbol token" (a numeric ID) for each stock. Some common ones:

Stock	Token
RELIANCE	2885
TCS	11536
INFY	1594
HDFCBANK	1333
ICICIBANK	4963
SBIN	3045
BHARTIARTL	10604
ITC	1660
KOTAKBANK	1922
HINDUNILVR	1394

For any other stock, check the AngelOne instrument list or search for it in your AngelOne trading app.

---

What's next

This is intentionally simple. Here's where you could take it:

• Option chain analysis — agent_with_options.py is a start, but you could build a dedicated options screener that flags unusual OI buildup or IV skew
• News sentiment — Use Claude's web search tool or feed in news headlines to gauge sentiment alongside technicals
• Automated alerts — Run the agent on a cron job, send yourself Telegram/Slack alerts when it flags something interesting
• Order placement — place_order.py has the scaffolding. Once you're confident in your strategy, flip DRY_RUN to False (carefully)
• Multi-agent setup — Build specialized agents (a screener agent, a risk monitor, a news agent) and orchestrate them together
• WebSocket streaming — SmartAPI supports real-time tick data via WebSocket — useful for intraday strategies
• Backtesting — Pull historical data, run your agent's analysis backwards, see how its calls would have played out

---

Disclaimer

This project is for educational purposes only. It is not financial advice.

• The analysis generated by Claude is based on historical price data and should not be used as the sole basis for trading decisions
• Always do your own research before making investment decisions
• The authors are not responsible for any financial losses incurred from using this code
• If you enable real order placement, you do so entirely at your own risk
• Past performance does not guarantee future results

---

License

MIT — do whatever you want with it.

---

Built by Anish. If you found this useful, star the repo.