dalal-mcp 0.1.0

Zero-config MCP server for Indian stock markets, with screening and market-breadth tools

dalal-mcp

A zero-config, free MCP server for Indian stock markets — with screening and market-breadth tools no other Indian-markets MCP server offers.

dalal-mcp lets any LLM client (Claude Desktop, Claude Code, OpenAI Codex, Cursor) query NSE/BSE equities in plain language. Unlike other Indian-markets MCP servers, it doesn't stop at quote lookup: it can run a stock screen and read market internals (StockBee-style breadth) — and it needs no API key.

"Which NIFTY stocks are up more than 4% today on above-average volume?"
"What's market breadth right now — how many stocks are above their 50-day average?"
"Give me Infosys fundamentals and its last month of prices."

Why this one

• 🆓 Zero-config & free. Uses Yahoo Finance data (.NS/.BO tickers). No signup, no paid API key — unlike servers that require a paid data plan.
• 🔎 Screening built in (screen_stocks) — filter a universe by price, % change, volume, and moving-average crosses. No other Indian-markets MCP server does this.
• 📊 Market breadth (market_breadth) — advancers/decliners, new 52-week highs/lows, % of stocks above their 50/200-day moving averages.
• 🧱 Reliable by design — thin data layer, fully unit-tested tools, graceful per-symbol error handling.

Install

uvx dalal-mcp

(Requires uv. No other setup.)

Configure your client

Claude Desktop / Claude Code (claude_desktop_config.json):

{
  "mcpServers": {
    "dalal": { "command": "uvx", "args": ["dalal-mcp"] }
  }
}

OpenAI Codex / Cursor — point your MCP config at the same command (uvx dalal-mcp, stdio transport).

Tools

Tool	What it does
search_symbol	Resolve a name/partial ticker to Indian-market symbols
get_quote	Latest price, day range, volume
get_historical	OHLCV candles over a period
get_fundamentals	P/E, market cap, EPS, sector, 52-wk range
screen_stocks	Filter a universe by numeric criteria
market_breadth	Advancers/decliners, new highs/lows, % above DMA
get_index	Index level + constituents (NIFTY 50, SENSEX)

Development

uv venv && uv pip install -e ".[dev]"
uv run pytest

Tests use mocked data — no network access required in CI.

License

MIT