# Flash Trade MCP

## Flash Trade MCP Server

Give AI agents direct access to Flash Trade — market data, trade previews, position management, and transaction building through natural language.

No cloning. No building. Follow the steps for your editor.

NPM: [`flash-trade-mcp`](https://www.npmjs.com/package/flash-trade-mcp) | GitHub: [`flash-trade/flash-trade-MCP`](https://github.com/flash-trade/flash-trade-MCP)

***

## Claude Code

Choose one of the two options below:

### Option A — Project-level (only available in one project)

1. `cd` into your project folder
2. Create `.mcp.json` in the project root

### Option B — Global (available in all projects)

1. Edit `~/.claude.json` (create it if it doesn't exist)

***

Then paste this block:

```json
{
  "mcpServers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

{% hint style="info" %}
If `~/.claude.json` already has a `mcpServers` object, add the `"flash-trade": { ... }` entry inside it rather than replacing the whole file.
{% endhint %}

Restart Claude Code — Flash Trade tools are now available.

***

### Claude Desktop (macOS)

The Connectors UI in Settings (where you see "Add custom connector") is for remote servers only. Local MCP servers like this one go in a JSON config file.

1. Open Claude Desktop → **Settings** → **Developer**
2. Click **"Edit Config"**
3. Find `claude_desktop_config.json` and open it with a text editor (right-click → Open With → TextEdit, or VS Code)
4. You will see something like this:

```json
{
  "preferences": {
    "coworkScheduledTasksEnabled": true,
    "sidebarMode": "task",
    ...other settings...
  }
}
```

5. Add a comma after the `preferences` closing brace, then paste the `mcpServers` block. The full file should look like:

```json
{
  "preferences": {
    "coworkScheduledTasksEnabled": true,
    "sidebarMode": "task",
    ...your existing settings stay unchanged...
  },
  "mcpServers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

6. Save the file
7. Go back to Claude Desktop. Click **Claude** in the menu bar → **Quit Claude** (or press Cmd+Q). Closing the window is not enough — you must fully quit the app
8. Reopen Claude Desktop
9. Start a new conversation. You should see a hammer icon in the chat input area — that means Flash Trade MCP is connected

***

### Claude Desktop (Windows)

1. Open Claude Desktop → **Settings** → **Developer**
2. Click **"Edit Config"**
3. Find `claude_desktop_config.json` and open it with Notepad or any text editor
4. You will see something like this:

```json
{
  "preferences": {
    "coworkScheduledTasksEnabled": true,
    "sidebarMode": "task",
    ...other settings...
  }
}
```

5. Add a comma after the `preferences` closing brace, then paste the `mcpServers` block. The full file should look like:

```json
{
  "preferences": {
    "coworkScheduledTasksEnabled": true,
    "sidebarMode": "task",
    ...your existing settings stay unchanged...
  },
  "mcpServers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

6. Save the file
7. Right-click Claude Desktop in the taskbar → **Close window**, or press Alt+F4. Make sure the app is fully closed, not just minimized
8. Reopen Claude Desktop
9. Start a new conversation. You should see a hammer icon in the chat input area — that means Flash Trade MCP is connected

***

### VS Code (GitHub Copilot)

Requires GitHub Copilot extension installed and an active Copilot subscription.

1. Open VS Code
2. Press **Cmd+Shift+P** (Mac) or **Ctrl+Shift+P** (Windows) to open the Command Palette
3. Type `MCP` and select **MCP: Add Server**
4. Choose **Workspace** (adds to this project) or **Global** (available in all projects)
5. This creates or opens a `mcp.json` file. Add the Flash Trade server inside the `servers` object:

```json
{
  "servers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

Note: VS Code uses `"servers"` not `"mcpServers"` — this is different from Claude.

6. Save the file
7. A **Start** button appears at the top of the file. Click it to start the MCP server
8. Open Copilot Chat (Ctrl+Alt+I / Cmd+Ctrl+I), select **Agent** mode, and Flash Trade tools are available

***

### Cursor

1. Open Cursor
2. Press **Cmd+Shift+P** (Mac) or **Ctrl+Shift+P** (Windows) to open the Command Palette
3. Type `MCP` and select **Cursor Settings: Open MCP Settings**
4. This opens a JSON file. Add the Flash Trade server inside the `mcpServers` object:

```json
{
  "mcpServers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

5. Save the file
6. Open the Command Palette again (Cmd/Ctrl+Shift+P), type `MCP` and select **Reload MCP Servers**
7. Flash Trade tools are now available in Cursor's AI chat

***

### Windsurf

1. Open Windsurf
2. Click the **Cascade** icon (the AI assistant panel)
3. Click the **hammer icon** at the top of the Cascade panel
4. Click **Configure** to open the MCP config file
5. Add the Flash Trade server inside the `mcpServers` object:

```json
{
  "mcpServers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

6. Save the file
7. Click the **Refresh** button next to the MCP server list, or restart Windsurf
8. Flash Trade should appear in the server list as connected

***

### Cline (VS Code Extension)

1. Open VS Code with the Cline extension installed
2. Click the **Cline** icon in the sidebar
3. Click the **MCP Servers** button (gear icon) in the Cline panel
4. Click **Configure MCP Servers** — this opens the Cline MCP settings file
5. Add the Flash Trade server:

```json
{
  "mcpServers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

6. Save the file
7. Back in the Cline MCP panel, Flash Trade should appear. Toggle it on if needed
8. Flash Trade tools are now available in Cline conversations

***

### OpenClaw

1. Open your OpenClaw config file at `~/.openclaw/openclaw.json`
2. Find the `mcpServers` object (or add one if it doesn't exist)
3. Add the Flash Trade server entry:

```json
{
  "mcpServers": {
    "flash-trade": {
      "command": "npx",
      "args": ["-y", "flash-trade-mcp"],
      "env": {
        "FLASH_API_URL": "https://flashapi.trade"
      }
    }
  }
}
```

If you already have other servers in `mcpServers`, add `"flash-trade": { ... }` as a new entry inside the existing object.

4. Save the file
5. Restart the gateway:

```bash
openclaw gateway restart
```

6. Flash Trade tools are now available to your OpenClaw agents

***

### ChatGPT

ChatGPT only supports **remote** MCP servers over HTTPS — it cannot run local stdio servers like `npx`. Flash Trade MCP currently uses stdio transport only.

**This means ChatGPT is not supported.**

***

### Claude.ai (Web)

Same limitation as ChatGPT — Claude.ai requires remote MCP servers with a URL. Flash Trade MCP runs locally via stdio.

**Not supported.** Use Claude Desktop or Claude Code instead.

***

***

Full tool reference and source code: [GitHub](https://github.com/flash-trade/flash-trade-MCP)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.flash.trade/flash-trade/flash-trade-protocol/build-on-flash/flash-trade-mcp.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.