TradingView MCP Market Data & Technical Analysis for AI Assistants

TradingView MCP server — real-time market data, technical indicators, screeners, and backtesting for Claude, ChatGPT, Cursor, Copilot, and any MCP client. Stocks, crypto, forex & futures across global exchanges. Backtesting + live sentiment + Yahoo Finance + 37 technical-analysis tools — the most complete TradingView MCP toolkit, all in one server.

🆕 The hosted server now draws. Ask for a chart and an interactive candlestick view (1d / 1h / 15m, optional Bollinger overlay) renders live inside the conversation via MCP Apps — and the AI reads the band values back to you:

Paste one URL (or pip install), then just ask. Real output, straight from the live server:

Try these first — they work in Claude, ChatGPT, Cursor, or any MCP client:

Show today's top crypto gainers on Binance
Run a full technical analysis of NVDA
What's the multi-timeframe read on gold?
Backtest an RSI strategy on BTC on the daily timeframe

[!NOTE] Independent open-source project — not affiliated with, endorsed by, or associated with TradingView Inc. "TradingView" is a trademark of its respective owner; this project consumes third-party market data and is not a TradingView product.

[!NOTE] Does it need — or risk — your TradingView account? No. This server does not log into, scrape, or automate a TradingView session, and it requires no TradingView account or API key. Market data is fetched server-side from public endpoints, so there is no account of yours in the loop and no browser/UI automation. (This is different from MCP servers that drive the TradingView Desktop app via Chrome DevTools.) You are responsible for ensuring your own use complies with the terms of any data source you point it at.

[!IMPORTANT] Not financial advice. Nothing produced by this software is investment, financial, legal, tax, or accounting advice. tradingview-mcp is an informational and educational analysis tool. Its outputs, including indicators, scores, signals, "trade setups", entries, stop losses, and targets, are computed from third party market data and are not recommendations to buy, sell, or hold any asset. It does not execute trades, manage money, or guarantee any result. Trading and investing carry a substantial risk of loss, and you can lose some or all of your capital. Always do your own research and consult a licensed professional before making any financial decision. You are solely responsible for your own decisions and for complying with the laws and regulations that apply to you. Market data may be delayed, inaccurate, or incomplete, and is provided without warranty.

[!TIP] Prefer zero setup? Use the hosted version. pro.cryptosieve.com serves all 37 tools as one connector URL for Claude.ai, ChatGPT, Copilot, and Cursor — no uv, pandas, or Python to wrangle. From $9/mo (Pro) or $29/mo (Pro+ — higher limits), with a 3-day free trial. Self-hosting stays free forever; hosted is just for folks who'd rather skip the ops. (Full self-host vs hosted comparison in Quick Start below.)

⭐ If this tool improves your workflow, please star the repo and consider sponsoring — it keeps the project alive and growing!

🎥 Framework Demo

https://github-production-user-asset-6210df.s3.amazonaws.com/67838093/478689497-4a605d98-43e8-49a6-8d3a-559315f6c01d.mp4

🆕 What's New

Stability & Strategy Expansion (May 2026)

• Async hot-path tools — 7 high-traffic tools (yahoo_price, stock_extended_hours, top_gainers, volume_breakout_scanner, multi_timeframe_analysis, financial_news, combined_analysis) converted to async def. FastMCP runs sync tools serialized on the event loop — async unlocks real intra-server parallelism so concurrent tool calls actually overlap. combined_analysis additionally fans its 3 sub-calls out via asyncio.gather for ~3× wall-clock improvement on the power tool. yahoo_price / stock_extended_hours use httpx.AsyncClient for true non-blocking I/O; sync-library tools (tradingview_ta, tradingview-screener, feedparser) are off-loaded via asyncio.to_thread. (PR — open)
• 9 backtest strategies (up from 6) — added rsi_pullback, keltner_breakout, and triple_ema, covering trend-pullback, ATR-normalized breakout, and SMA200-filtered EMA cross edges. compare_strategies now ranks the full 9.
• Resilience layer — automatic retry + 60-second TTL cache on the TradingView screener provider, eliminating transient "Expecting value" errors on combined_analysis and multi_timeframe_analysis. (PR #32 — merged)
• Financial news service rebuild — replaces deprecated Reuters RSS endpoints with Yahoo Finance, MarketWatch, and CNBC. Fixes the long-standing count: 0 bug on financial_news. (PR #33 — merged)
• TA throttle — caps concurrent tradingview_ta calls (default 4) + min 0.8s spacing between starts. Prevents parallel bursts of combined_analysis / multi_timeframe_analysis from hitting TradingView's empty-body rate-limit cliff. Tunable via env vars. (PR #34 — merged)
• Walk-forward backtesting (walk_forward_backtest_strategy) — train/test split with overfitting verdict (ROBUST / MODERATE / WEAK / OVERFITTED).
• Hourly (1h) timeframe support across backtest_strategy, compare_strategies, and walk_forward_backtest_strategy.
• Full trade log + equity curve outputs (include_trade_log=True, include_equity_curve=True).

🏗️ Architecture

✨ Why tradingview-mcp?

🚀 Quick Start (5 Minutes)

Two ways to run it — the same 37 tools either way:

☁️ Zero setup: paste one connector URL into Claude.ai, ChatGPT, Copilot, or Cursor → start a 3-day free trial. Everything below is for self-hosting.

Install via pip

pip install tradingview-mcp-server

Optional: news & sentiment (free Marketaux key)

financial_news and market_sentiment (plus the news/sentiment parts of combined_analysis) are powered by Marketaux — licensed market news with per-entity sentiment. Grab a free API key (100 requests/day; the server caches for 4h and shares one fetch between news and sentiment, so the free tier goes a long way) and set:

export MARKETAUX_API_TOKEN=your_token_here   # optional

Without a token those two tools return a friendly "not configured" note — all other tools work normally.

Claude Desktop Config (claude_desktop_config.json)

Note: On macOS, GUI apps like Claude Desktop may not have ~/.local/bin in their PATH. Use the full path to uvx to avoid "command not found" errors.

{
  "mcpServers": {
    "tradingview": {
      "command": "/Users/YOUR_USERNAME/.local/bin/uvx",
      "args": ["--from", "tradingview-mcp-server", "tradingview-mcp"]
    }
  }
}

On Linux, replace /Users/YOUR_USERNAME with /home/YOUR_USERNAME. On Windows, use %USERPROFILE%\.local\bin\uvx.exe.

Codex Plugin Config

This repository also includes mcp-only Codex plugin metadata:

• .codex-plugin/plugin.json
• .codex-mcp.json

The plugin uses the same PyPI package entrypoint:

{
  "mcpServers": {
    "tradingview": {
      "command": "uvx",
      "args": ["--from", "tradingview-mcp-server", "tradingview-mcp"]
    }
  }
}

After installing or enabling the Codex plugin, restart Codex so the MCP server is loaded in the next session. Depending on your Codex version, codex mcp list may show registered MCP servers, but tool availability should be verified in a fresh Codex session.

Or run from source

git clone https://github.com/atilaahmettaner/tradingview-mcp
cd tradingview-mcp
uv run tradingview-mcp

🛠️ Troubleshooting

🪟 Windows: MCP error -32001: Request timed out on first launch

Symptom — you see this in the Claude Desktop logs shortly after adding the config:

[tradingview] Server started and connected successfully
[tradingview] Message from client: initialize ...
[60 seconds later]
[tradingview] notifications/cancelled — reason: "MCP error -32001: Request timed out"

Why it happens: Python 3.14 is not supported yet. uvx downloads tradingview-mcp-server, creates a fresh virtualenv, and installs dependencies the first time it runs. Some native dependencies in the MCP stack do not currently publish compatible Python 3.14 wheels, so installation can fall back to source builds or fail before Claude Desktop finishes initializing the server.

Fix — pin to Python 3.13 (has prebuilt pandas wheels):

{
  "mcpServers": {
    "tradingview": {
      "command": "uvx",
      "args": ["--python", "3.13", "--from", "tradingview-mcp-server", "tradingview-mcp"]
    }
  }
}

On macOS use the full path to uvx (see the note in Quick Start). On Windows uvx is typically %USERPROFILE%\.local\bin\uvx.exe.

Alternative — pre-install once, then let Claude Desktop reuse the cache:

# Run in a terminal before launching Claude Desktop
uv tool install --python 3.13 tradingview-mcp-server

After the install finishes, start Claude Desktop with the normal config and the server will come up instantly (cache is already warm).

Credit: @wyh4444 for the original report in #24.

⚠️ Error Envelope Format

Tools that have adopted the structured error format return either their normal payload or an error envelope:

{"error": {"code": "ALL_BATCHES_FAILED", "message": "All 5 batches failed; first error: JSONDecodeError(...)", "batches_attempted": 5, "batches_failed": 5, "first_error": "...", "retryable": true}}

Why: the previous [] / {"error": "Analysis failed: ..."} strings made it impossible to distinguish "no matches today" from "upstream rate-limit cliff." The new envelope is programmatically branchable by code, and retryable tells the caller whether waiting and retrying can help (upstream storms pass; a missing dependency won't fix itself).

Full coverage (every failure mode — upstream outage, empty symbol list, missing dependency, unexpected exception — returns an envelope, never a raw string/traceback): top_gainers, top_losers, bollinger_scan, rating_filter, volume_breakout_scanner, smart_volume_scanner.

Partial adoption (main error paths return envelopes): coin_analysis, volume_confirmation_analysis, consecutive_candles_scan, futures_market_overview, futures_top_movers, stock_prices, and the backtest input guards. Remaining tools migrate in batches — tracked in #76.

Detecting an error:

result = volume_breakout_scanner(exchange="KUCOIN")
if isinstance(result, dict) and "error" in result:
    code = result["error"]["code"]
    if result["error"].get("retryable"):
        # Wait + retry, raise alert, fall back to single-batch call, etc.
        ...
else:
    for row in result:
        ...

Stable codes are defined in core/errors.py.

📱 Use via Telegram, WhatsApp & More (OpenClaw)

Connect this server to Telegram, WhatsApp, Discord and 20+ messaging platforms using OpenClaw — a self-hosted AI gateway. Tested & verified on Hetzner VPS (Ubuntu 24.04).

How It Works

OpenClaw routes Telegram messages to an AI agent. The agent uses trading.py — a thin Python wrapper — to call tradingview-mcp functions and return formatted results. No MCP protocol needed between OpenClaw and the server; it's a direct Python import.

Telegram → OpenClaw agent (AI model) → trading.py (bash) → tradingview-mcp → Yahoo Finance

Quick Setup

# 1. Install UV and tradingview-mcp
curl -LsSf https://astral.sh/uv/install.sh | sh && source ~/.bashrc
uv tool install tradingview-mcp-server

# 2. Configure OpenClaw channels
cat > ~/.openclaw/openclaw.json << 'EOF'
{
  channels: {
    telegram: {
      botToken: "YOUR_BOT_TOKEN_HERE",
    },
  },
}
EOF

# 3. Configure gateway + agent
openclaw config set gateway.mode local
openclaw config set acp.defaultAgent main

# 4. Set your AI model (choose ONE option below)
openclaw configure --section model

# 5. Install the skill + tool wrapper
mkdir -p ~/.agents/skills/tradingview-mcp ~/.openclaw/tools
curl -fsSL https://raw.githubusercontent.com/atilaahmettaner/tradingview-mcp/main/openclaw/SKILL.md \
  -o ~/.agents/skills/tradingview-mcp/SKILL.md
curl -fsSL https://raw.githubusercontent.com/atilaahmettaner/tradingview-mcp/main/openclaw/trading.py \
  -o ~/.openclaw/tools/trading.py && chmod +x ~/.openclaw/tools/trading.py

# 6. Start the gateway
openclaw gateway install
systemctl --user start openclaw-gateway.service

Choose Your AI Model

OpenRouter is not required — use whichever provider you have a key for:

# Examples — set your chosen model:
openclaw config set agents.defaults.model "openrouter/google/gemini-3-flash-preview"  # via OpenRouter
openclaw config set agents.defaults.model "anthropic/claude-sonnet-4-5"               # Anthropic direct
openclaw config set agents.defaults.model "google/gemini-2.5-flash"                   # Google direct

⚠️ Important: Prefix must match your provider. google/... needs a Google API key. openrouter/... needs an OpenRouter key.

⚠️ Common Mistakes

Test Your Bot

Once running, send your Telegram bot:

market snapshot
backtest RSI strategy for AAPL, 1 year
compare all strategies for BTC-USD

👉 Full OpenClaw Setup Guide →

Unlike basic screeners, this framework deploys specialized AI agents that debate findings in real-time:

1. 🛠️ Technical Analyst — Bollinger Bands (±3 proprietary rating), RSI, MACD
2. 🌊 Sentiment & Momentum Analyst — Reddit community sentiment + price momentum
3. 🛡️ Risk Manager — Volatility, drawdown risk, mean-reversion signals

Output: STRONG BUY / BUY / HOLD / SELL / STRONG SELL with confidence score

🔧 All 37 MCP Tools

📊 Backtesting Engine

9 Strategies to Test:

• rsi — RSI oversold/overbought mean reversion
• bollinger — Bollinger Band mean reversion
• macd — MACD golden/death cross
• ema_cross — EMA 20/50 Golden/Death Cross
• supertrend — ATR-based Supertrend trend following 🔥
• donchian — Donchian Channel breakout (Turtle Trader style)
• rsi_pullback — Dip-buy in confirmed uptrend (SMA50>SMA200 + RSI<40 entry) 🆕
• keltner_breakout — ATR-normalized breakout (EMA20 + 2·ATR upper band) 🆕
• triple_ema — EMA 20/50 cross gated by SMA200 trend filter 🆕

🆕 strategies require period='1y' or '2y' so the SMA200 trend filter can complete its warmup.

Metrics you get: Win Rate, Total Return, Sharpe Ratio, Calmar Ratio, Max Drawdown, Profit Factor, Expectancy, Best/Worst Trade, vs Buy-and-Hold, with realistic commission + slippage simulation.

Example prompt: "Compare all 9 strategies on MSFT for 2 years"
→ #1 triple_ema:        +15.1% | Sharpe:  0.0 | WR: 100%
→ #2 keltner_breakout:  +14.3% | Sharpe:  4.7 | WR:  40%
→ #3 bollinger:         +12.2% | Sharpe:  4.1 | WR:  64%
→ Buy & Hold:            -2.1%

💰 Yahoo Finance — Real-Time Prices (New in v0.6.0)

Supports: Stocks (AAPL, TSLA, NVDA), Crypto (BTC-USD, ETH-USD, SOL-USD), ETFs (SPY, QQQ, GLD), Indices (^GSPC, ^DJI, ^IXIC, ^VIX), FX (EURUSD=X), Turkish (THYAO.IS, SASA.IS)

🌍 Global Stock Screener — Common & Preferred Shares

🧠 AI Sentiment & Intelligence

📈 Technical Analysis Core

🌍 Multi-Exchange Support

💬 Example AI Conversations

You: "Give me a full market snapshot right now"
AI: [market_snapshot] → S&P500 -3.4%, BTC +0.1%, VIX 31 (+13%), EUR/USD 1.15

You: "What is Reddit saying about NVDA?"
AI: [market_sentiment] → Strongly Bullish (0.41) | 23 posts | 18 bullish

You: "Backtest RSI strategy on BTC-USD for 2 years"
AI: [backtest_strategy] → +31.5% return | 100% win rate | 2 trades | B&H: -5%

You: "Which of the 9 strategies worked best on MSFT in the last 2 years?"
AI: [compare_strategies] → triple_ema #1 (+15.1%, WR 100%), keltner_breakout #2 (+14.3%), macd last (-23.4%)

You: "Run walk-forward backtest on supertrend for SPY"
AI: [walk_forward_backtest_strategy] → Verdict: ROBUST (avg robustness 0.92) | OOS return +8.5%

You: "Analyze TSLA with all signals: technical + sentiment + news"
AI: [combined_analysis] → BUY (Technical STRONG BUY + Bullish Reddit + Positive news)

💖 Support the Project

This framework is free and open source, built in spare time. If it saves you hours of research or helps you make better decisions, please consider sponsoring:

Every sponsor directly funds new features like Walk-Forward Backtesting, Twitter/X sentiment, and managed cloud hosting.

📋 Roadmap

• TradingView technical analysis (30+ indicators)
• Multi-exchange screener (Binance, KuCoin, MEXC, EGX, US stocks)
• Reddit sentiment analysis
• Live financial news (Yahoo / MarketWatch / CNBC / CoinDesk / CoinTelegraph)
• Yahoo Finance real-time prices
• Backtesting engine (9 strategies + Sharpe / Calmar / Expectancy)
• Walk-forward backtesting (overfitting detection)
• Resilience layer (retry + TTL cache) on screener provider
• Hourly (1h) backtesting timeframe
• Twitter/X market sentiment
• Paper trading simulation
• Managed cloud hosting (no local setup)

📄 License

MIT License — see LICENSE for details.

Disclaimer: This tool is for educational and research purposes only. It does not constitute financial advice. Always do your own research before making investment decisions.