How to Use TradingView MCP with Claude AI: Full Setup Guide

If you’ve been looking for a way to connect Claude AI to your TradingView charts in real time, TradingView MCP makes that possible today. Instead of pasting screenshots into a chat window, Claude reads your live chart data directly — every candle, every indicator value, every price tick — and responds to plain English commands. Furthermore, with a tool like PickMyTrade, you can take that AI-generated signal and execute it as a live trade automatically, turning an analytical workflow into a full end-to-end trading system.

In this guide, you’ll learn exactly what TradingView MCP is, how to set it up using the LewisWJackson tradingview-mcp-jackson repository, what you can do with it, and how to automate the trades it generates.

What Is TradingView MCP?

TradingView MCP is an open-source MCP (Model Context Protocol) server that connects Claude Code to your locally running TradingView Desktop application. Specifically, it works through the Chrome DevTools Protocol (CDP) — a standard debugging interface built into all Chromium-based applications, including TradingView Desktop, which runs on Electron.

In practical terms, this means Claude does not read a screenshot or a delayed API snapshot. Instead, Claude reads the underlying live chart data the same way a browser’s developer console reads a webpage. As a result, Claude can see your exact indicator values, candle data, watchlist, and open positions — in real time.

The Model Context Protocol itself is an open standard developed by Anthropic that allows AI assistants to connect to external tools and data sources through a standardized interface. Consequently, any MCP server you configure in Claude Code or Claude Desktop can provide Claude with live context from that system.

Why Use the LewisWJackson Fork?

Several TradingView MCP implementations exist. However, the LewisWJackson/tradingview-mcp-jackson fork stands out for traders who want a production-ready, daily-driver workflow rather than a proof of concept.

Specifically, this fork adds:

• Morning brief workflow — scans every symbol in your watchlist, reads all indicator values, applies your personal rules (rules.json), and prints a bias summary (bullish/bearish/neutral) with key levels to watch
• One-shot setup prompt — copy a single prompt into Claude Code and it configures the whole server for you, replacing the manual multi-step process
• Fixed TradingView Desktop v2.14+ launch bug — the auto-launch feature (tv_launch) correctly detects and starts the desktop app on current versions
• CLI access — every MCP tool is also accessible as a tv CLI command, so you can script workflows outside Claude

Moreover, the repo is actively maintained and has accumulated significant adoption from the trading community, making it a reliable foundation.

Prerequisites

Before you set up TradingView MCP, make sure you have the following in place:

• TradingView Desktop app installed (not the browser version) — this is required because CDP only works with the desktop Electron app
• A valid TradingView subscription — the tool connects to your existing account and does not bypass any access controls
• Claude Code installed (npm install -g @anthropic-ai/claude-code) — this is Anthropic’s CLI for running Claude with MCP servers
• Node.js 18+ installed on your machine
• Git installed

How to Set Up TradingView MCP with Claude AI

Step 1: Clone the Repository

Open your terminal and run the following commands:

bash

git clone https://github.com/LewisWJackson/tradingview-mcp-jackson.git ~/tradingview-mcp-jackson
cd ~/tradingview-mcp-jackson
npm install

This clones the repository to your home directory and installs all dependencies.

Step 2: Add the MCP Server to Claude Code

Next, you need to register the MCP server in your Claude Code configuration. Open (or create) the file ~/.claude/.mcp.json and add the following entry:

json

{
  "mcpServers": {
    "tradingview": {
      "command": "node",
      "args": ["/Users/YOUR_USERNAME/tradingview-mcp-jackson/server.js"]
    }
  }
}

Replace YOUR_USERNAME with your actual system username. Additionally, if you prefer a project-level config rather than a global one, you can place .mcp.json at the root of your working project folder instead.

Step 3: Configure Your Trading Rules

The morning brief workflow is powered by a personal rules.json file that you define. Copy the example to get started:

bash

cp rules.example.json rules.json

Open rules.json and fill in:

• watchlist — the ticker symbols you actively trade (e.g., ["BTCUSDT", "EURUSD", "SPY"])
• bias_criteria — the indicator conditions that make a symbol bullish or bearish for you
• risk_rules — maximum position size, risk per trade, or any filters you want applied

Notably, this file is local to your machine and is never transmitted externally. All data processing happens locally.

Step 4: Launch TradingView with CDP Enabled

TradingView Desktop must be running with the Chrome DevTools Protocol port open. You have two options:

Option A — Use the tv_launch tool (recommended): Simply ask Claude to launch TradingView:

Ask Claude: "Launch TradingView for me"

Claude will use the tv_launch MCP tool to auto-detect and start the app with the correct flags.

Option B — Launch manually: Find the TradingView Desktop executable and launch it with the debug flag:

bash

# macOS example
open -a TradingView --args --remote-debugging-port=9222

Step 5: Verify the Connection

Open Claude Code and run a health check:

Ask Claude: "Run tv_health_check"

A successful connection returns:

json

{
  "success": true,
  "cdp_connected": true
}

If cdp_connected is false, TradingView is not running with the debug port enabled. Go back to Step 4 and try Option B manually.

What You Can Do with TradingView MCP and Claude

Once connected, the combination of Claude AI and TradingView MCP opens up a substantial number of practical workflows.

Read Live Chart Data

You can ask Claude to read your current chart state in plain English:

• “What is the RSI value on the 1-hour BTCUSDT chart?”
• “What indicators do I have on my current chart?”
• “Show me the last 20 candles for EURUSD on the daily”

Claude reads the actual values directly from your running TradingView session, not from a static data source.

Write and Debug Pine Script

Furthermore, TradingView MCP enables a full AI-assisted Pine Script development loop:

1. You describe what you want: “Write a Pine Script strategy that goes long when RSI crosses above 40 from oversold”
2. Claude writes the script, injects it into TradingView, compiles it, reads any errors, fixes them, and recompiles — automatically
3. You end up with a working, error-free script in your chart

This is particularly powerful for traders who are not developers but want custom indicators or strategies.

Run Your Morning Brief

The morning brief command is arguably the most practical daily-driver feature. Once your rules.json is configured, you simply ask:

"Run my morning brief"

Claude scans every symbol in your watchlist, reads all indicator values, applies your rules, and returns a structured summary like:

BTCUSDT   → BULLISH  | RSI: 58 | EMA cross: above | Key level: 68,400
EURUSD    → NEUTRAL  | RSI: 49 | No clear setup
SPY       → BEARISH  | RSI: 34 | Below 200 EMA | Key level: 512

Consequently, you start each trading session with a clear, rule-based view of your entire watchlist — in seconds.

Multi-Pane Chart Layouts

In addition, you can command Claude to set up multi-symbol chart layouts:

• “Set up a 4-pane layout with BTC, ETH, SOL, and XRP on the 4-hour”
• “Switch to the daily chart and add VWAP and Bollinger Bands”

This eliminates the need to click through menus entirely.

How to Automate Trades Generated by TradingView MCP

Reading charts with AI is powerful. However, the real leverage comes from closing the loop between AI analysis and live trade execution. This is where PickMyTrade fits into the workflow.

What PickMyTrade Does

PickMyTrade is a cloud-based trading automation platform that connects TradingView alerts directly to live broker accounts. It supports over 10 brokers including Tradovate, Rithmic, Interactive Brokers, TradeStation, and TradeLocker, as well as proprietary trading firms like Apex Trader Funding and Topstep.

• No coding required
• Sub-200ms execution speed
• Unlimited trades and strategies for a flat $50/month
• 24/7 cloud operation — alerts execute even when your computer is off

The Combined Workflow

You can combine TradingView MCP + Claude + PickMyTrade into a seamless system:

1. Claude reads your charts via TradingView MCP and identifies setups matching your rules
2. Claude writes or updates a Pine Script strategy in TradingView that fires alerts on those conditions
3. PickMyTrade receives the TradingView alerts and executes the trades on your broker account automatically

In short, Claude handles the analysis and strategy logic, TradingView generates the alerts, and PickMyTrade executes the trades. The result is a fully automated trading pipeline driven by AI analysis of live chart data.

TradingView MCP vs. Manual Trading: What Changes

It is worth understanding what this technology actually changes in your workflow:

Workflow Step	Without MCP	With TradingView MCP + Claude
Reading indicators	Click through charts manually	Ask Claude in plain English
Pine Script writing	Code manually or Google examples	Claude writes, injects, debugs automatically
Morning scan	Open each chart one by one	morning_brief scans full watchlist in seconds
Strategy refinement	Trial and error in TradingView	Describe the change, Claude implements it
Trade execution	Place trades manually	PickMyTrade executes alerts automatically

Overall, the combination reduces the routine, repetitive parts of a trader’s workflow and frees up time for the higher-level decisions that actually require human judgment.

Important Limitations and Caveats

TradingView MCP is a powerful tool, but there are several important points to understand before relying on it:

• Requires TradingView Desktop — the browser version of TradingView does not support CDP and therefore does not work with this setup
• Requires a TradingView subscription — the tool connects to your existing account; it does not provide free data access
• Local only — all processing happens on your machine; the MCP server does not transmit your chart data to any external service
• Not financial advice — Claude’s chart analysis reflects the patterns in your indicators and rules, not a guarantee of trading profitability. Always apply your own risk management

Frequently Asked Questions

What is TradingView MCP and how does it work?

TradingView MCP is an open-source MCP (Model Context Protocol) server that connects Claude AI to your TradingView Desktop application using the Chrome DevTools Protocol (CDP). CDP is a standard debugging interface built into Chromium-based apps. TradingView Desktop runs on the Electron framework, which is Chromium-based, so opening a debug port allows Claude to read live chart data — candles, indicator values, watchlists — in real time without screen capture.

Do I need a TradingView paid subscription to use TradingView MCP?

Yes, a valid TradingView subscription is required. The MCP server connects to your locally running TradingView Desktop session and uses the data you already have access to through your account. It does not bypass any TradingView paywall or provide access to data beyond your subscription tier.

What is the difference between TradingView MCP and the LewisWJackson fork?

The original tradingview-mcp project was the first open-source implementation connecting Claude Code to TradingView Desktop. The LewisWJackson/tradingview-mcp-jackson fork builds on that foundation and adds a morning brief workflow, a personal rules.json config system, a one-shot setup prompt, and a fix for the TradingView Desktop v2.14+ launch bug. For most traders, the Jackson fork is the better starting point in 2026.

Can I use TradingView MCP with Claude Desktop instead of Claude Code?

Yes, TradingView MCP can be configured with Claude Desktop as well as Claude Code. The MCP configuration approach is the same you add the server entry to your MCP config file. However, Claude Code is generally recommended for trading workflows because it offers a persistent terminal environment better suited to running commands and scripts alongside chart analysis.

How does PickMyTrade fit into a TradingView MCP workflow?

PickMyTrade handles the trade execution side of the workflow. Specifically, Claude via TradingView MCP handles analysis and Pine Script generation, while TradingView generates alerts based on those scripts. PickMyTrade then receives those alerts and places the actual trades on your broker account in under 200 milliseconds. Together, these tools create a complete pipeline from AI analysis to live execution without you needing to manually click “buy” or “sell.”

Is TradingView MCP safe to use? Does it share my trading data?

TradingView MCP processes all data locally on your machine. No chart data, indicator values, or account information is transmitted to any external server by the MCP tool itself. The connection is between your local TradingView Desktop app and your local Claude Code instance only. As with any open-source tool, you should review the source code at github.com/LewisWJackson/tradingview-mcp-jackson before use.

Conclusion

TradingView MCP with Claude AI represents a meaningful shift in how traders interact with their charts. Rather than manually clicking through indicators and charts, you can use plain English to read data, build strategies, and run daily scans all powered by a live connection to your TradingView Desktop session. The LewisWJackson fork is the most practical starting point for traders who want a daily-driver workflow rather than a prototype.

To close the loop from AI analysis to automated execution, connect your TradingView alerts to PickMyTrade a platform trusted by over 10,000 traders to execute alerts on Tradovate, Rithmic, Interactive Brokers, and more, with sub-200ms speed and no coding required.

Ready to automate? Get started with PickMyTrade free →

You may also like:

Claude AI Pine Script: Write V6 Strategies That Actually Compile
DeerFlow AI Agent: How to Use for Trading Strategy Research
Automate News Trading: 2026 Calendar Guide
Diversified Bot Strategy: The 3-Bot Portfolio for 2026 Profits

External references:

• LewisWJackson/tradingview-mcp-jackson on GitHub
• Model Context Protocol — Anthropic
• PickMyTrade — Automate TradingView Alerts