Short answer

You can connect your Arrow account to AI assistants like Claude, Cursor, and Windsurf completely free of charge using the new Arrow Model Context Protocol (MCP). This gives you instant access to a personal financial assistant that understands your portfolio and can answer complex questions in any language of your choice.

What is Model Context Protocol (MCP)?

Model Context Protocol (MCP) is a communication standard that enables AI assistants to interact with real-world data and services. Rather than relying only on pre-trained knowledge, AI tools can now access live information from your Arrow account.

Similar to Arrow Connect APIs, which are now free, you can access your Arrow account through MCP without any additional charges. To learn more about Arrow Connect, visit our Arrow Connect guide.

MCP enables AI assistants to provide:

Real-time data access: Access current market prices, not just historical data

Personalised insights: Analyse your specific portfolio, not just generic insights

Interactive capabilities: Retrieve precise information you need, not just answer questions

Natural conversation: Ask about complex market data in plain language

Without MCP, AI assistants can only access generic information available on the internet and cannot give responses specific to your account.

How to set up Arrow MCP

Claude (Desktop)

Install Node.js

Download and install Node.js from nodejs.org.

Verify your installation by opening the command prompt and typing node –version

Configure Claude Desktop

Open your Claude Desktop application.

Go to Settings (gear icon).

Click on Developer in the left sidebar.

Click Edit Config.

Add the following configuration:

Save and restart Claude Desktop.

For a visual walkthrough of the setup process, watch this step-by-step video guide on configuring MCP for Claude Desktop.

Verify your connection

In Claude Desktop, look for the tools/hammer icon in the chat interface.

Click it to verify Arrow MCP tools are available.

Follow the authorisation prompts to connect to your Arrow account.

Visual Studio Code

You can also configure Arrow MCP if you prefer using Visual Studio Code with AI extensions:

Prerequisites

Visual Studio Code installed

Node.js installed

VS Code GitHub Copilot extension or another AI extension that supports MCP

Configuration steps

Open VS Code settings (File > Preferences > Settings, or press Ctrl+,).

Search for “copilot chat mcp” or navigate to the GitHub Copilot Chat configuration.

Click on Edit in settings.json.

Add the following configuration to your settings.json file:

Save the settings file and restart VS Code

Open the Copilot Chat panel and use the /mcp command to verify that Arrow appears as an available MCP server.

When prompted, authorise your Arrow account to connect with VS Code.

For more detailed information on setting up MCP servers in VS Code, refer to the official documentation.

Setup on Linux (Ubuntu/Debian)

You can use two unofficial builds for Claude Desktop on Linux:

Using aaddrick’s Debian/Ubuntu build

Visit the GitHub Repository.

Install using these commands:

Visit the GitHub Repository.

Run quickly with:

Create the configuration directory and file:

Add the following configuration:

Install Claude Code:

When prompted, set up the Arrow MCP server with the remote URL

What can you do with Arrow MCP?

With Arrow MCP connected to Claude, you can have conversations about your portfolio and the markets in natural language. Here are the currently available features:

Portfolio analysis

You can ask for:

A breakdown of your portfolio performance across different time periods

Insights into your sector exposure and diversification

Alerts when your portfolio drifts from target allocations

Information about which of your holdings has the highest percentage gain today

Your current positions and their unrealised P&L

Research on stocks in natural language while seeing how they would fit with your existing portfolio

Portfolio reports, comparisons, and visualisations instead of manual calculations

Research capabilities

You can:

Set up GTT orders to implement systematic entry and exit strategies at predetermined price points

Analyse your portfolio performance to identify patterns and improve your approach

Plan your trades, refine your strategies, size your positions and manage risk

Create custom technical indicators and overlay them with price charts

Pull historical data to backtest strategies and visualise their performance

Analyse correlations between your holdings and market indices

Get current price data, trading volumes, and key statistics for any stock

Account information

You can check:

Your available margins and buying power

Your account’s P&L for the day

Your portfolio’s risk profile

Please note: Order placement, historical trade data, portfolio data, and some other features are currently unavailable. Arrow MCP primarily provides market data and portfolio analysis capabilities at this time.

Security and privacy

iRage takes security seriously with the Arrow MCP implementation:

Limited access: Claude can only access data you explicitly authorise

No credential storage: Your iRage credentials never pass through Claude. Instead, authentication happens externally through Arrow’s secure two-factor authentication flow

Read-only operations: All operations are read-only – order placement is not available except for GTT orders