Indian Stocks Mcp

indian-stocks-mcp

Ask Claude about Indian stocks. Minute-by-minute. No setup.

Connect one hosted MCP server and query Indian market history from conversation: 136 NSE index series, 117 NIFTY 100 stock symbols, 6+ years of OHLCV candles, and backtests with 8 strategy types and a full Indian cost model.

[](https://vjaiswal-nifty-mcp.hf.space) [](https://vjaiswal-nifty-mcp.hf.space/mcp) [](LICENSE)

Hosted endpoint: https://vjaiswal-nifty-mcp.hf.space/mcp

Transport: streamable-http

---

Connect

Claude Desktop

Add this to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "indian-stocks": {
      "type": "http",
      "url": "https://vjaiswal-nifty-mcp.hf.space/mcp"
    }
  }
}

Restart Claude Desktop after editing the file.

Claude.ai

Go to Settings -> Integrations -> Add custom MCP, then paste:

https://vjaiswal-nifty-mcp.hf.space/mcp

VS Code, Cursor, and other MCP clients

Use the same hosted HTTP endpoint:

{
  "mcpServers": {
    "indian-stocks": {
      "type": "http",
      "url": "https://vjaiswal-nifty-mcp.hf.space/mcp"
    }
  }
}

For clients that ask for transport explicitly, choose streamable-http.

---

What to ask

> "What was RELIANCE's intraday price action on Union Budget day 2024?"

> "Run a 20/50 SMA crossover backtest on NIFTY 50 daily from 2019 to 2024. What's the Sharpe ratio?"

> "Compare RSI mean reversion vs EMA crossover on INFY 1-minute data. Which had better risk-adjusted returns?"

> "Show me NIFTY BANK 15-minute candles during the March 2020 crash week"

> "Backtest a Supertrend strategy on NIFTY IT with a 2% trailing stop loss. What was the worst drawdown?"

> "Run a MACD crossover backtest on RELIANCE from 2020 to 2024. Compare it against Bollinger Bands on the same data."

> "Use a Stochastic oscillator strategy on BANK NIFTY 15-minute data with a 1.5% stop loss"

> "Which NIFTY 100 stocks have more than 5 years of 1-minute data available?"

> "Compare 5 different SMA window combinations on NIFTY 50 and rank them by profit factor"

> "Show me the latest day's 1-minute candles for NIFTY 50"

> "What's the average 1-minute range for HDFC BANK in Q1 2024?"

Claude calls the MCP tools automatically, so answers are grounded in the hosted data instead of guessed from model memory.

---

Available Data

| Dataset | Symbols | Intervals | History | Notes | |---------|---------|-----------|---------|-------| | Indices | 136 index series | 1minute, 5minute, 15minute, 60minute, day | 6+ years | Includes broad-market, sector, factor, bond, thematic, and volatility series | | Stocks | 117 NIFTY 100 stock symbols as advertised by the hosted server | 1minute | 6+ years | Use list_symbols(dataset="stocks") for the live authoritative list |

Row scale: 187M+ candles.

Timezone: all timestamps are returned in Asia/Kolkata (+05:30).

Refresh cadence: twice-weekly (Monday and Thursday) from Kaggle-backed source datasets.

Representative index symbols: NIFTY 50, NIFTY BANK, NIFTY IT, NIFTY PHARMA, NIFTY AUTO, NIFTY REALTY, INDIA VIX.

Representative stock symbols: RELIANCE, TCS, INFY, HDFCBANK, ICICIBANK, WIPRO, BAJFINANCE, TITAN, MARUTI, COALINDIA.

See [docs/symbols.md](docs/symbols.md) for the static public symbol catalog and notes on verification.

---

MCP Tools

| Tool | What Claude can do | |------|--------------------| | dataset_status | Check data availability, latest available date, row counts per dataset | | list_symbols | Browse symbols, intervals, date coverage, and per-symbol row counts | | get_candles | Fetch raw OHLCV candles for a symbol, interval, date range, and row limit. Use desc=True to get the latest N candles without needing to know the date | | run_backtest | Run a strategy config and return metrics, trades, and equity curve | | compare_strategies | Run up to 10 strategy configs side-by-side | | get_backtest_report | Retrieve a saved backtest report by run_id, or list recent reports | | help | List all available tools with descriptions, parameters, and usage examples |

All tools are read-only. This MCP does not place trades, manage portfolios, or write to brokerage accounts.

---

Backtesting

Strategy types

| Strategy | What it does | Key parameters | |----------|-------------|----------------| | sma_crossover | Buy when fast SMA crosses above slow SMA | fast_window, slow_window | | ema_crossover | Buy when fast EMA crosses above slow EMA | fast_window, slow_window | | rsi_mean_reversion | Buy oversold, exit overbought | rsi_window, oversold, overbought | | breakout | Buy on N-period high breakout | lookback_window | | supertrend | ATR-based trend following | atr_window, multiplier | | macd_crossover | Buy when MACD line crosses above signal line | fast_window (12), slow_window (26), macd_signal (9) | | bollinger_bands | Buy at lower band, sell at upper band | bb_window (20), num_std (2.0) | | stochastic | Buy oversold %K/%D, sell overbought | stoch_window (14), stoch_signal (3), oversold (20), overbought (80) |

Exit rules

Every strategy can include:

{
  "stop_loss_pct": 0.02,
  "take_profit_pct": 0.05,
  "trailing_stop_pct": 0.03
}

Indian cost model

Backtests can include:

STT, brokerage per order, SEBI charges, exchange transaction fee, stamp duty, GST, and slippage.

Metrics

Backtests return final equity, total return, CAGR, Sharpe ratio, max drawdown, win rate, trade count, average trade return, total costs, full trades, and full equity curve.

---

Example Backtest Config

Ask Claude to run this, or pass it directly to run_backtest:

{
  "name": "NIFTY 50 MACD crossover",
  "dataset": "index",
  "symbol": "NIFTY 50",
  "interval": "day",
  "start": "2019-01-01",
  "end": "2024-12-31",
  "initial_cash": 100000,
  "strategy": {
    "type": "macd_crossover",
    "fast_window": 12,
    "slow_window": 26,
    "macd_signal": 9,
    "stop_loss_pct": 0.03
  },
  "position_sizing": {
    "type": "fractional",
    "fraction": 1.0
  },
  "costs": {
    "brokerage_per_order": 20,
    "stt_sell_pct": 0.001,
    "slippage_pct": 0.0005
  }
}

---

Troubleshooting

Check hosted server health:

curl https://vjaiswal-nifty-mcp.hf.space/health

A healthy response looks like:

{"status":"ready","mode":"remote","db_ready":true}

If your client cannot connect:

• Confirm the URL is exactly https://vjaiswal-nifty-mcp.hf.space/mcp.
• Confirm the MCP type is http or the transport is streamable-http.
• Restart the MCP client after changing config.
• If HuggingFace is waking the Space, retry after a short wait.

---

FAQ

Is this free?

Yes. The hosted server is free to use and does not require an API key.

How fresh is the data?

The source datasets are refreshed twice a week (Monday and Thursday).

Does this provide live prices?

No. It provides historical OHLCV candles. Live quotes are out of scope.

Can it trade for me?

No. Every MCP tool is read-only.

Something broken or a symbol missing?

Open an issue in this repository and include the symbol, interval, date range, and MCP client you used.

---

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: vishalmdi
• Source: vishalmdi/indian-stocks-mcp
• License: MIT
• Homepage: https://vjaiswal-nifty-mcp.hf.space

Install and usage instructions live in the source repository linked above.