nse-bse-indian-stock-market-data-mcp

What is this?

Tapetide MCP Server is a Model Context Protocol server that connects AI assistants to real-time Indian stock market data. It covers all ~8,200 stocks listed on NSE and BSE — from large-cap Nifty 50 to SME stocks.

Ask your AI to look up any stock, run a screener with 326 fundamental filters or real-time technical indicators, pull quarterly financials, check analyst consensus ratings, track your portfolio P&L, monitor FII/DII institutional flows, or get today's bulk deals — all through natural language.

Compatible with: Claude Desktop, Claude Code, ChatGPT, Cursor, Windsurf, Kiro, VS Code, Codex, Zed, Gemini, Grok, OpenCode, Antigravity, and any MCP-compatible client.

Quick Start

Option 1: Remote MCP (No install — claude.ai, chatgpt.com, Grok, Gemini)

Add this URL directly in your AI chat app:

https://mcp.tapetide.com/mcp

Authentication happens automatically via Google OAuth. No token needed.

Option 2: Remote MCP with Token (Claude Code, VS Code, Kiro, Zed)

For code editors that support URL-based MCP servers with custom headers:

1. Get a free token at tapetide.com/settings/tokens
2. Add to your MCP config:

{
  "mcpServers": {
    "tapetide": {
      "type": "url",
      "url": "https://mcp.tapetide.com/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_TOKEN_HERE"
      }
    }
  }
}

Option 3: Local MCP via npm (Claude Code, Codex, Cursor, Windsurf, VS Code, Gemini CLI, Kiro, OpenCode)

For stdio-based MCP clients. No cloning or building required — runs via npx:

1. Get a free token at tapetide.com/settings/tokens
2. Add to your MCP config:

{
  "mcpServers": {
    "tapetide": {
      "command": "npx",
      "args": ["-y", "tapetide-mcp"],
      "env": {
        "TAPETIDE_TOKEN": "your_token_here"
      }
    }
  }
}

Node.js 18+ required for the local option. Run node --version to check.

How It Works

┌─────────────────┐     stdio (JSON-RPC)     ┌──────────────────┐     HTTPS      ┌─────────────────────┐
│  AI Assistant   │ ◄──────────────────────► │  tapetide-mcp    │ ◄────────────► │  mcp.tapetide.com   │
│  (Claude, etc.) │                           │  (npm package)   │                │  (Cloudflare Worker) │
└─────────────────┘                           └──────────────────┘                └─────────────────────┘

The npm package is a lightweight stdio bridge with zero runtime dependencies. It:

• Reads JSON-RPC from stdin, forwards to the remote Tapetide MCP server, writes responses to stdout
• Auto-detects framing: Content-Length (VS Code, Claude Desktop) or newline-delimited JSON (Kiro, Claude Code)
• Exchanges your refresh token for a 1-hour HMAC access token, auto-refreshes before expiry
• Handles SSE responses from the remote server
• Identifies your MCP client to the remote via User-Agent, and echoes the negotiated protocol version
• Warns on stderr when the server rate-limits a call, naming the retry delay
• Replays the Mcp-Session-Id the server mints at initialize on every later request, so the remote can tie a session's calls together
• Forwards MCP Apps traffic untouched — resources/list, resources/read and the _meta.ui hints on get_price_history, market_heatmap and screen_stocks reach your client exactly as the server sends them

All 55 tools and their logic run on the remote server — the npm package is just the transport layer. Because it forwards JSON-RPC verbatim, tools shipped on the remote are available immediately without upgrading this package.

Authentication

Generate a free personal token at tapetide.com/settings/tokens. Works for both remote and local MCP.

Tools

Tip: ask your assistant to call read_me first. It returns the full in-session guide — every tool by category, usage patterns, and the rules the server expects clients to follow.

Three tools carry an interactive view for hosts that support MCP Apps (Claude, Claude Desktop, VS Code Copilot): get_price_history renders the price and volume chart, market_heatmap the index treemap, and screen_stocks the sortable results table. Text-only hosts get the same JSON result with nothing missing.

🔍 Discovery & Screening (5 tools)

📊 Company Analysis (9 tools)

🏛️ Market-Wide Data (11 tools)