tradingview-mcp-thaqif
A local MCP server that connects TradingView Desktop to AI assistants via Chrome DevTools Protocol, enabling Pine Script development, multi-timeframe analysis, and Wyckoff scanning with a rules engine.
README
tradingview-mcp-thaqif
Drive TradingView Desktop from Claude (and other MCP clients) for Pine Script v5/v6, FCPO/futures, and Wyckoff multi-timeframe analysis.
A local MCP server that bridges TradingView Desktop to an AI assistant over the Chrome DevTools Protocol, shipped with a 64-rule rules.json that constrains how the AI writes Pine and reads the market. MIT-licensed, runs entirely on your machine.
Improved fork of tradingview-mcp-jackson by Lewis W. Jackson, itself built on the original tradingview-mcp by @tradesdontlie. Full credit to both — see NOTICE.
[!WARNING] Not affiliated with TradingView Inc. or Anthropic. This tool connects to the TradingView Desktop app already running on your machine via the Chrome DevTools Protocol. It requires your own valid TradingView subscription and does not bypass any paywall. It surfaces analysis only — it does not place trades. Read the Disclaimer and SECURITY.md before use.
[!NOTE] All processing is local. No TradingView data leaves your machine; the server has no telemetry.
What it does
- Read your chart — symbol, timeframe, indicator values, and custom Pine output (lines, labels, tables, boxes).
- Develop Pine Script v5/v6 — inject source, smart-compile, and read errors in a loop.
- Manage alerts, replays, and screenshots.
morning_brief— scan a watchlist, read every indicator, and return structured data so the model produces a Wyckoff multi-timeframe bias, driven by yourrules.json.
What's new in this fork
| Feature | What it does |
|---|---|
rules.json v2.0.0 |
A 64-rule engine across 9 categories (Pine v5/v6, indicators, multi-timeframe, Wyckoff, FCPO, backtesting, alerts, risk, MCP tools). See docs/RULESET.md. |
examples/fcpo_wyckoff_mtf.pine |
A worked, rule-compliant Pine v6 indicator (Spring/UTAD, non-repaint HTF, FCPO sessions). |
npm + .mcpb packaging |
Install via npx, or one-click in Claude Desktop via a Desktop Extension bundle. |
| Inherited from the jackson fork | morning_brief, session_save/session_get, tv CLI, and the tv_launch fix for TradingView Desktop v2.14+. |
Requirements
- Node.js 18+
- TradingView Desktop running with the Chrome DevTools Protocol enabled (see SETUP_GUIDE.md)
- A valid TradingView subscription
Install
Full per-client snippets (Cursor, Cline, VS Code, local clone) are in the install section below.
Claude Desktop — one click: install the .mcpb from Settings -> Extensions, or build it from manifest.json (see DISTRIBUTION.md).
Claude Desktop — config file (claude_desktop_config.json):
{
"mcpServers": {
"tradingview": {
"command": "npx",
"args": ["-y", "tradingview-mcp-thaqif"]
}
}
}
Cursor (~/.cursor/mcp.json), Cline (cline_mcp_settings.json), and VS Code (.vscode/mcp.json) use the same command/args. Run from a local clone with "command": "node", "args": ["/abs/path/tradingview-mcp-thaqif/src/server.js"].
Then, with TradingView Desktop open, verify with the tv_health_check tool.
The rules engine
rules.json is loaded as AI context so the assistant treats critical/high rules as hard constraints. The 9 categories: pine_script, indicators, multi_timeframe, wyckoff, fcpo, backtesting, alerts, risk_management, mcp_tools. Each rule has an id, rationale, severity, and a runnable example. Full guide: docs/RULESET.md. Worked example: examples/.
Publishing / distribution
This repo is packaged for npm, the official MCP Registry (server.json), and a Claude .mcpb bundle (manifest.json). Step-by-step in DISTRIBUTION.md.
Disclaimer
For research and educational use. Nothing here is financial advice. Trading FCPO, crude oil, and crypto futures involves substantial risk of loss. This tool is not affiliated with TradingView Inc. or Anthropic, PBC; you are responsible for complying with TradingView's Terms of Use. The server never executes trades — surface signals and place trades yourself.
Attribution & license
MIT — see LICENSE and the full attribution chain in NOTICE. Original by @tradesdontlie; fork foundation by @LewisWJackson.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。