# Vibe Coding with the eToro MCP Server

How to use the eToro MCP server with Cursor, Claude, and other AI code editors to build trading apps without writing API calls by hand.

---


You don't need to dig through scattered notes to wire up the eToro API. With the **eToro MCP (Model Context Protocol) server**, your AI editor can read the same official documentation you would use by hand—endpoints, schemas, and examples—while you stay in your project.

This is **vibe coding** for fintech.

## What Is MCP?

The Model Context Protocol is an open standard that lets AI assistants (like Claude in Cursor, Antigravity, or VS Code) call external tools. Instead of pasting guesses into your codebase, you describe what you want to build, and the assistant pulls structured answers from connected servers.

The eToro MCP server gives your AI assistant access to the full API documentation — endpoints, schemas, authentication patterns, and code examples. Your assistant queries the docs server to write accurate API code instead of guessing.

## Setting Up

### 1. Add the MCP server

The eToro API docs MCP is hosted at the portal — no `npx` package. Add this to your editor’s MCP config (e.g. Cursor: `.cursor/mcp.json`):

```json skip-test
{
  "mcpServers": {
    "etoro-api-docs": {
      "url": "https://api-portal.etoro.com/mcp"
    }
  }
}
```

### 2. Sign in at the API portal

Use [api-portal.etoro.com](https://api-portal.etoro.com) to register your app and get API keys when you call the REST API directly. The MCP connection uses the hosted docs server at that portal.

### 3. Start vibing

Open your editor and start asking. The assistant can search and read the official API portal documentation through the MCP connection so answers match what the portal publishes.

## Example Workflows

### "Show me the top 5 gaining stocks today"

Ask your assistant to find the right Market Data or instrument endpoints, query parameters, and response shapes in the docs, then sketch the HTTP calls or client code. You still run requests against the real API with your own keys; the MCP layer helps the assistant stay aligned with documented behavior instead of inventing endpoints.

### "Build me a portfolio tracker dashboard"

Describe the UI and data you need—for example a React view with P&L, allocation, and a sortable positions table. The assistant can look up portfolio-related endpoints, required headers, and pagination from the documentation, then generate components and types that match those contracts.

### "Place a demo trade"

Ask the assistant to read the Trading API documentation for order payloads, demo vs. real accounts, and authentication before you implement. That cuts down on mismatches between sample code and the live API.

## Tips for Better Results

1. **Be specific** — "Get the price of Bitcoin" works better than "show me crypto"
2. **Mention demo** — Always specify "demo account" when testing trades
3. **Build incrementally** — Start with data fetching, then add trading logic
4. **Check the types** — Ask the AI to show you the response schema before building UI

## Supported Editors

| Editor | MCP Support | Status |
|--------|-------------|--------|
| Cursor | Native | Fully supported |
| Claude Desktop | Native | Fully supported |
| VS Code + Continue | Plugin | Supported |
| Antigravity | Native | Supported |
| Zed | Plugin | Beta |

## Get Started

1. Visit [api-portal.etoro.com](https://api-portal.etoro.com/vibe-code/cursor) for setup instructions
2. Get your API key from the [API Portal](https://api-portal.etoro.com)
3. Start building with your favorite AI editor

The future of fintech development is conversational. Start vibing with eToro today.