TradingView MCP connects AI assistants such as Claude Code to your TradingView Desktop app, so an agent can read your charts, indicators and watchlists and help analyze markets in your own workflow. It exposes TradingView’s state to AI through the Model Context Protocol. The project has about 6,900 stars on GitHub and was updated in 2026.
The server bridges your locally running TradingView Desktop and an MCP-compatible AI client. Through it, the assistant can pull the current chart’s symbol and timeframe, read indicator values and study outputs, inspect watchlists and drawings, and return structured analysis grounded in what is actually on your screen โ rather than the model guessing prices from memory. Typical workflows include asking the agent to summarize the technical setup across several timeframes, check whether your indicators agree before a trade, or batch-review a watchlist and flag the charts worth a closer look. Everything runs locally against your own TradingView session: no credentials are shared with third parties, and the tool is aimed at personal workflow automation for discretionary traders who already live in TradingView. Setup involves running the MCP server alongside TradingView Desktop and registering it with your AI client.
TradingView MCP is free and open source. You need your own TradingView account (and subscription, if you use premium indicators) and your own AI assistant.
TradingView MCP is for traders who chart in TradingView Desktop and want an AI assistant that can actually see their charts and indicators while helping with analysis.