dev-tools 4 min read

TradingView MCP: AI-Assisted Chart Analysis for Claude Code

tradingview-mcp is a Model Context Protocol server that connects Claude Code to your TradingView Desktop for personal chart analysis workflows.

By
Share: X in
GitHub social preview card for tradesdontlie/tradingview-mcp

TL;DR: tradesdontlie/tradingview-mcp is a JavaScript Model Context Protocol server that connects Claude Code to TradingView Desktop for personal chart-analysis workflows. 6,479 stars on GitHub, single-purpose tool.

What it is

tradesdontlie/tradingview-mcp describes itself as “AI-assisted TradingView chart analysis — connect Claude Code to your TradingView Desktop for personal workflow automation.” The product is an MCP server that runs locally, talks to TradingView Desktop on the user’s machine, and exposes the chart data through a tool Claude Code can call. The repository is written in JavaScript, sits at 6,479 stars, and the GitHub API does not report a recognized SPDX license identifier — the API field returns NOASSERTION, which means the repository has not declared a license that GitHub’s license database can normalize. The homepage field is empty, so the repository itself is the product surface. The GitHub topic list is also empty in the API response, which is unusual but not a deal-breaker. The repository was created on 2026-03-29, which makes it about six months old.

The shape is the standard MCP-server pattern: a process you start on your machine, a manifest Claude Code reads, and a set of tool calls that wrap TradingView Desktop’s automation hooks.

What it does well

The interesting question for any “Claude Code + X” tool is what makes it worth wiring up versus just opening TradingView in another window. The description and the architecture point at three answers.

  • Personal workflow automation. The description explicitly says “personal workflow automation” rather than “team trading desk,” which sets expectations — this is a tool for an individual trader who wants Claude Code to read their charts and answer questions.
  • MCP-native integration. The repository name ends in -mcp, which is the standard convention for Model Context Protocol servers. Claude Code discovers it through the same channel it discovers any other MCP tool.
  • TradingView Desktop as the data source. TradingView Desktop exposes automation hooks that the MCP server wraps, so the charts the agent reads are the ones on the user’s screen rather than a parallel feed.
  • JavaScript footprint. The JavaScript language field means the install path is npm install plus a one-line MCP config, not a Python virtualenv.
  • Small surface area. A single-purpose tool with a short description is easier to reason about than a generic charting library, which matters when you are wiring it into an agent loop.

The 6,479-star count at 79 stars today reflects a tool that has found a niche audience rather than going viral.

How to try it

The fastest path is the standard MCP install flow, which Claude Code picks up the moment the server is on the path.

# Install the server
git clone https://github.com/tradesdontlie/tradingview-mcp
cd tradingview-mcp
npm install
npm run build

# Register it with Claude Code's MCP config
claude mcp add tradingview node $(pwd)/build/index.js

Once registered, Claude Code can answer chart questions directly.

> Open the SPY 4-hour chart and explain the last two swing highs.

When it fits, and when it doesn’t

TradingView MCP is the right tool for an individual trader who already lives in Claude Code and wants the agent to read charts on the same screen they are watching. It is a strong fit for personal journaling, for “explain this setup” questions on the fly, and for any workflow where the bottleneck is asking a question of your charts rather than drawing them. It is less attractive for serious backtesting that needs historical data APIs, for multi-user trading desks that need access control and audit trails, or for traders who prefer TradingView’s own Pine Script over an MCP-mediated agent loop. The unspecified license is worth a sanity check before you embed it into a commercial workflow.

The 6,479-star count, the six-month-old repository, and the single-purpose positioning are the signals worth weighing: a focused tool that solves one specific problem and does not pretend to be more.

Source and Accuracy Notes

  • Project page: https://github.com/tradesdontlie/tradingview-mcp
  • Source repository: tradesdontlie/tradingview-mcp (verified via GitHub API html_url)
  • License: NOASSERTION (GitHub API could not normalize a SPDX id — see repo for license terms)
  • Stars at time of writing: 6,479
  • Language: JavaScript
  • Topics: none returned by GitHub API
  • Homepage: not set in GitHub API response
  • Created: 2026-03-29