tradingview-mcp

Product Introduction

1. Definition: The TradingView MCP Bridge is an open-source Model Context Protocol (MCP) server. Technically, it is a middleware application that establishes a bidirectional communication channel between an AI assistant (specifically Claude Code) and the locally installed TradingView Desktop application.
2. Core Value Proposition: It exists to solve the problem of manual, repetitive chart analysis by enabling AI-assisted automation of personal trading workflows. Its primary value is bridging the gap between advanced large language models (LLMs) and professional financial charting software, allowing for automated technical analysis, Pine Script development, and research process streamlining directly from live charts.

Main Features

1. AI-Powered Chart Reading & Analysis: The tool provides a suite of MCP tools that allow an AI agent to programmatically read data from a live TradingView chart. This includes fetching real-time quotes, OHLCV data, current values from technical indicators (like RSI, MACD), and extracting outputs from custom Pine Script indicators such as drawn lines, labels, and data tables. It works by connecting to TradingView's Electron app via the Chrome DevTools Protocol (CDP) on localhost:9222 to execute JavaScript in the app's context and scrape the DOM.
2. Programmatic Chart Control & Automation: Users can command the AI to perform actions on their chart. This includes changing symbols and timeframes, adding or removing indicators, adjusting indicator settings, navigating to specific dates, managing multi-pane layouts and tabs, and even automating UI interactions like clicks and keyboard inputs. This is achieved through the same CDP connection, simulating user interactions and API calls to TradingView's internal application programming interfaces.
3. Integrated Pine Script Development Cycle: This feature creates a closed-loop environment for Pine Script coding. The AI can write a script, inject it directly into the TradingView Pine Editor, compile it, read back any compilation errors or console log outputs, and iterate based on the results. It supports both client-side compilation (via the chart) and server-side syntax checking, significantly accelerating the indicator and strategy development process.
4. Local Data Streaming & Monitoring: The tool includes a CLI with tv stream commands that poll the local TradingView Desktop instance and output deduplicated data as JSON Lines (JSONL). This enables the creation of local monitoring scripts, dashboards, or data loggers that can track price changes, indicator values, or Pine Script outputs in real-time, with all processing occurring on the user's machine without external data transmission.

Problems Solved

1. Pain Point: Manual and time-consuming chart analysis. Traders and analysts often perform repetitive tasks like scanning multiple symbols, recording indicator readings, or manually testing Pine Script code, which is inefficient and prone to oversight.
2. Target Audience: The primary users are retail traders, quantitative analysts, and algorithmic trading developers who use TradingView for technical analysis and Pine Script for strategy development. A secondary audience includes researchers studying human-AI collaboration in financial workflows.
3. Use Cases: Essential scenarios include: rapidly prototyping and debugging Pine Script indicators with AI assistance; conducting multi-symbol or multi-timeframe analysis by automating chart navigation; backtesting trading ideas by using the AI to control the chart replay mode; and automating the collection of specific data points (e.g., daily support/resistance levels) from charts for personal journals or models.

Unique Advantages

1. Differentiation: Unlike generic screen-scraping bots or unofficial TradingView APIs, this tool operates strictly through the official, local Desktop application using a standard debugging protocol (CDP). It does not require API keys, reverse-engineer network traffic, or modify TradingView's software. Compared to manual operation, it provides a structured, programmatic interface where every function is also available as a pipe-friendly JSON CLI command.
2. Key Innovation: Its core innovation is the application of the Model Context Protocol (MCP) as a bridge between LLMs and stateful desktop applications. It translates natural language requests from an AI into precise, low-level CDP commands to interact with TradingView's complex GUI and internal state, and then structures the graphical output back into legible, context-aware data for the AI. This turns the charting platform into a "tool" the AI can use.

Frequently Asked Questions (FAQ)

1. Is the TradingView MCP tool safe and legal to use? The tool is safe from a security perspective as it runs locally and connects only to your own TradingView app via a debug port you explicitly enable. Legally, you are responsible for ensuring your usage complies with TradingView's Terms of Use, which prohibit automated data collection and scraping. This tool facilitates such interaction, so use it at your own risk regarding your TradingView account standing.
2. Do I need a TradingView paid subscription to use this MCP server? Yes, a valid TradingView subscription (Pro, Pro+, or Premium) is required to run the TradingView Desktop application with real-time market data. This MCP tool does not bypass any paywalls; it merely interacts with the application you are already running and are authorized to use.
3. Can this MCP server execute automated trades on my brokerage account? No, this tool cannot execute real trades. It is designed for chart interaction, analysis, and Pine Script development only. It controls the TradingView application interface, not your brokerage connection. Any "trades" placed in Replay mode or the Strategy Tester are simulated.
4. What happens if TradingView updates their Desktop application? The tool relies on TradingView's internal, undocumented Electron structure, which may change with any update. This can cause features to break. The project documentation recommends pinning your TradingView Desktop version if stability for automation is critical to your workflow.
5. How does the TradingView MCP Bridge handle real-time data streaming? All data streaming occurs locally. The tv stream commands poll your running TradingView Desktop app via the local Chrome DevTools Protocol connection. The data is processed, deduplicated, and output to your terminal or script. No market data is sent to external servers by this tool.