Tapetide MCP Server — Stock Market MCP Server for AI Stock Analysis

Search & Discovery

• search_stocks: Find stocks by name, symbol, BSE code, or ISIN. Filter by sector and industry.
• get_trending_stocks: Top gainers, losers, and high-volume stocks from Nifty 500.

Company Data

• get_company_profile: Sector, about, pros/cons, fundamentals, growth metrics, quote. Optionally include technicals, ratings, peers.
• get_stock_events: News (sentiment-tagged), corporate actions (dividends, splits, bonuses), and company filings.
• get_stock_ownership: Dividend history and mutual fund holdings for a stock.

Quotes & Prices

• get_stock_quote: Current price, change, volume, market cap, PE, PB, 52-week range.
• get_batch_quotes: Quotes for up to 20 stocks in one call.
• get_price_history: Daily or weekly OHLCV with delivery %. Up to 2,000 days of history.

Financials & Analysis

• get_financials: P&L, balance sheet, cash flow, ratios — quarterly and annual.
• get_shareholding: Promoter, FII, DII, and public holdings over time.
• get_forecasts: Analyst EPS, revenue, EBITDA estimates with actuals vs estimates comparison.

Stock Screener

• screen_stocks: Custom fundamental screener with 326 ratios — valuation, profitability, growth, balance sheet, cash flow, shareholding, plus cross-field and technical filters.
• screen_stocks_technical: Technical screener using real-time indicators — RSI, MACD, moving-average crossovers, Bollinger Bands, volume, and momentum.
• get_screener_ratios: Search or list the 326 financial ratios available to the screener so you can build precise queries.

Market Insights

• get_market_pulse: Today's FII/DII flows, market valuations, and top technical signals in one call.
• get_fii_dii_detail: Institutional flows — daily, weekly, monthly, yearly with F&O participant positioning.
• get_fpi_sectors: FPI sector-wise AUM, fortnight change, and 1-year cumulative flows.
• get_market_news: Latest market-wide news across categories — sentiment-tagged, with related stock symbols.
• market_valuations: Index PE/PB/DY over time for Nifty 50, Bank Nifty, Nifty IT, and more.
• market_signals: Technical signals — breakouts, crossovers, volume spikes, RSI extremes.
• market_heatmap: Constituent stocks with price changes and metrics for 16 indices.

Market Data

• market_deals: Bulk and block deals with client name, quantity, and price.
• market_fno_ban: F&O ban list and stocks approaching ban (MWPL ≥ 80%).
• market_deliveries: Stocks with highest delivery percentage (≥ 50%).
• market_ipo: Current and upcoming IPOs with subscription data.
• market_mtf: Margin Trading Facility — funded positions and trends.
• market_slbm: Stock Lending & Borrowing with yield calculations.

What is the Tapetide MCP server?

Tapetide MCP is a Model Context Protocol server that connects AI assistants to Indian stock market data. Available as a remote server (URL-based, Google OAuth or personal token) and a local npm package (token-based) — both with the same free-tier limits. It provides 27 market-data tools (plus watchlist and portfolio tools once you connect your account) covering stock search, company profiles, financials, shareholding, price charts, technicals, analyst ratings, forecasts, a stock screener with 326 ratios, FII/DII flows, bulk deals, F&O data, IPOs, MTF, SLBM, market valuations, signals, heatmaps, and more — for all NSE and BSE listed stocks.

What is Model Context Protocol (MCP)?

MCP is an open standard created by Anthropic that lets AI assistants connect to external data sources and tools. Think of it as a universal plug for AI — one protocol that works across Claude, ChatGPT, Cursor, VS Code, and other AI tools. Instead of each AI building custom integrations, MCP provides a single standard way to connect. Tapetide implements this protocol to give your AI assistant access to Indian stock market data.

How does authentication work?

Tapetide offers three auth methods. Remote MCP (via URL) uses Google OAuth — just sign in with your Google account when prompted, no keys needed. Remote + Token — use a personal API token as a Bearer header, no browser sign-in needed (great for code editors). Local MCP (via npm) uses API tokens via the TAPETIDE_TOKEN environment variable. Personal tokens work for both remote and local MCP, and the free-tier limits are identical no matter which method you pick. Get a free token at tapetide.com/settings/tokens. All methods are free.

Is it free?

Yes. Tapetide MCP is completely free. The free tier gives you 50 MCP tool calls per day (up to 1,000 per month), and the limits are identical across every connection method — OAuth, personal token, and local npm. Just sign in with Google or grab a free API token to get started. Need more? The Starter (400/day) and Pro (1,200/day) plans are on the pricing page.

What AI tools are supported?

Tapetide MCP works with Claude.ai, Claude Desktop, ChatGPT (Plus/Pro/Team/Enterprise/Edu), Grok, Cursor, Windsurf, VS Code (GitHub Copilot), Kiro, Antigravity, Zed, Gemini CLI, Codex, OpenCode, Command Code, Cline, and any MCP-compatible client. Web apps (Claude.ai, ChatGPT, Grok) connect directly via their connector UI. Code editors and CLI tools can use either the remote URL (OAuth or a personal token) or the local npm package (npx -y tapetide-mcp with an API token).

Can I use Tapetide MCP with Claude.ai?

Yes. Go to Claude.ai → Customize → Connectors → + → Add custom connector. Enter the server URL (https://mcp.tapetide.com/mcp), sign in with Google when prompted, and your Tapetide tools will be available in any chat. Works on Free (1 connector), Pro, Max, Team, and Enterprise plans.

Can I use Tapetide MCP with ChatGPT?

Yes. ChatGPT supports remote MCP servers through Connectors. Go to Settings → Connectors → Create, enter the Tapetide MCP server URL (https://mcp.tapetide.com/mcp), and sign in with Google when prompted. Available on Plus, Pro, Team, Enterprise, and Edu plans.

Can I use Tapetide MCP with Cursor?

Yes. Add the server URL to a .cursor/mcp.json file, then open Settings → MCP and click Connect to sign in with Google (OAuth — no token needed). Prefer not to sign in via browser? Add your personal token as an Authorization: Bearer header instead. See the Cursor setup above for the exact config.

Can I use Tapetide MCP with VS Code?

Yes. Two options: (1) Remote — create a .vscode/mcp.json file with the URL config, VS Code natively supports remote MCP URLs. (2) Local — create a .vscode/mcp.json file with the npx tapetide-mcp config and your API token. See the setup instructions above for exact configs.

Can I use Tapetide MCP with Gemini?

Yes, through Gemini CLI. Add the local config (npx -y tapetide-mcp with your TAPETIDE_TOKEN, from a free token at tapetide.com/settings/tokens) to your ~/.gemini/settings.json file. The Gemini web app does not yet have a built-in MCP connector UI.

Can I use Tapetide MCP with Grok?

Yes. Grok supports custom MCP servers through Connectors. Go to grok.com/connectors → New Connector → Custom, paste the Tapetide MCP server URL (https://mcp.tapetide.com/mcp), click Add Connector, and sign in with Google when the auth page opens. Your Tapetide tools then work in any Grok chat.

Can I use Tapetide MCP with Codex, OpenCode, or Command Code?

Yes, all three connect via the remote MCP server with OAuth. Codex: run "codex mcp add tapetide --url https://mcp.tapetide.com/mcp" then "codex mcp login tapetide". OpenCode: run "opencode mcp add" (choose Remote and paste the URL) or add a remote entry to opencode.json. Command Code: run "cmd mcp add --transport http tapetide https://mcp.tapetide.com/mcp". Each opens a browser for Google sign-in, then type /mcp to verify. See the CLI Tools tabs in the setup section above for exact steps.

Can I use Tapetide MCP with DeepSeek?

DeepSeek does not currently have a native MCP client in its web interface. You can use DeepSeek models through MCP-compatible clients like Cursor, VS Code, or OpenCode and connect Tapetide MCP there. As MCP adoption grows, we expect more platforms to add native support.

Do I need Node.js installed?

Only for the local stdio bridge (npx -y tapetide-mcp), which needs Node.js 18+ and is used by clients that can't connect to a remote URL, like Claude Desktop. The remote MCP needs no Node.js: web apps (Claude.ai, ChatGPT, Grok) connect through their connector UI, and editors and CLIs (Cursor, VS Code, Kiro, Claude Code, Codex, OpenCode, Antigravity, Zed) connect via the URL with a Google sign-in.

What data is available?

Tapetide covers ~8,200 stocks listed on NSE and BSE. You get: real-time quotes, quarterly and annual financials (P&L, balance sheet, cash flow, ratios), shareholding patterns (promoter, FII, DII, public), historical OHLCV price data up to 2,000 days, 20+ technical indicators (RSI, SMA, EMA, MACD, Bollinger Bands, ADX, ATR, Stochastic, MFI, CCI, Williams %R, Supertrend, pivot points), analyst consensus ratings and target prices, EPS/revenue/EBITDA forecasts, a fundamental stock screener with 326 ratios plus a real-time technical screener, FII/DII institutional flows with F&O participant positioning, FPI sector-wise data, bulk/block deals, F&O ban list, IPOs with subscription data, MTF, SLBM, market valuations (PE/PB/DY for major indices), technical signals, stock heatmaps for 16 indices, market news, corporate actions, dividends, and mutual fund holdings.

What tools are available?

The market-data tools are organized into 7 categories: Search & Discovery (search_stocks, get_trending_stocks), Company Data (get_company_profile, get_stock_events, get_stock_ownership), Quotes & Prices (get_stock_quote, get_batch_quotes, get_price_history), Financials & Analysis (get_financials, get_shareholding, get_forecasts), Stock Screener (screen_stocks, screen_stocks_technical, get_screener_ratios), Market Insights (get_market_pulse, get_fii_dii_detail, get_fpi_sectors, get_market_news, market_valuations, market_signals, market_heatmap), and Market Data (market_deals, market_fno_ban, market_deliveries, market_ipo, market_mtf, market_slbm). Once you connect your account you also get watchlist and portfolio tools (add, update, remove, and read). Your AI assistant picks the right tool automatically based on your question.

How does the stock screener work?

There are three screener tools. screen_stocks runs fundamental screens against 326 financial ratios — valuation (PE, PB, market cap, EV/EBITDA), profitability (ROE, ROCE, margins), growth (sales/profit growth over 1/3/5 years), balance sheet (debt/equity, current ratio), cash flow, shareholding (promoter/FII/DII holding), and historical returns — combined with AND/OR logic and cross-field comparisons like "Current price > DMA 200". screen_stocks_technical runs real-time technical screens — RSI, MACD, moving-average crossovers (e.g. golden cross), Bollinger Bands, volume spikes, and momentum. And get_screener_ratios lets you search the full ratio catalog so you can build precise queries.

How is this different from using the Tapetide website directly?

The MCP server gives your AI assistant direct access to the same data that powers tapetide.com, but through natural conversation. Instead of navigating pages and clicking through tabs, you just ask questions in plain English. Your AI can also combine data from multiple tools in a single response — for example, comparing financials of two companies while checking their technical indicators and analyst ratings simultaneously.

Is the data real-time?

Quotes are near real-time (delayed by a few minutes during market hours). Financials, shareholding, and other fundamental data are updated daily after market close. Market data like FII/DII flows, bulk deals, and F&O ban lists are updated as soon as exchanges publish them. Historical OHLCV data goes back up to 2,000 trading days.

What are the rate limits?

The free tier allows 50 MCP tool calls per day, up to 1,000 per month — the same across every connection method (OAuth, personal token, and local npm). A light per-minute fair-use cap smooths bursts but a normal session never hits it. Each tool call counts as one request; batch operations (like get_batch_quotes for 20 stocks) count as a single request. Rate limit headers (X-RateLimit-Reset, Retry-After) are included so your client knows when the window resets. Need higher limits? Starter raises it to 400/day and Pro to 1,200/day.

Is this the best MCP server for stock research?

Tapetide is the most comprehensive stock market MCP server available for Indian markets. With 27+ market-data tools covering all NSE and BSE stocks, it offers more depth than any other financial MCP server — including a fundamental screener with 326 ratios, a real-time technical screener, FII/DII institutional flow tracking, 20+ technical indicators, analyst forecasts, and real-time market data. Available as both remote MCP (Google OAuth or personal token) and a local npm package (API token). Free to use.

Can I use this for algorithmic trading or automated stock research?

Yes. The MCP server provides all the data you need for automated equity research — real-time quotes, historical OHLCV data up to 2,000 days, technical indicators, screener with 100+ filters, and market signals. You can build AI agents that screen stocks, analyze financials, track institutional flows, and generate research reports. The stock market API integration works through any MCP-compatible AI agent framework.

How does this compare to using a stock market API directly?

Traditional stock market APIs require you to write code, handle authentication, parse responses, and build UIs. With Tapetide MCP, your AI assistant handles all of that — you just ask questions in natural language. The AI automatically picks the right tools, combines data from multiple endpoints, and presents the analysis in a readable format. It is the easiest way to integrate Indian stock market data into any AI workflow.

Can I analyze Nifty 50, Bank Nifty, and other indices?

Yes. The market_valuations tool provides PE, PB, and dividend yield data over time for Nifty 50, Nifty 500, Bank Nifty, Nifty IT, Nifty Midcap 50, and Nifty Next 50. The market_heatmap tool shows all constituent stocks for 16 indices with price changes, market cap, and sector data. You can also screen stocks within specific indices using the stock screener.

Does it work with machine learning and LLM stock prediction workflows?

Yes. The get_price_history tool provides up to 2,000 days of daily OHLCV data with delivery percentage — ideal for training ML models. Combined with technical indicators, fundamentals, and shareholding data, you have a complete feature set for machine learning stock prediction. The MCP protocol makes it easy to integrate this data into any LLM-based analysis pipeline.