Canva MCP

How to Install and Use Canva MCP

Canva MCP is a tool that helps your AI assistant work smoothly with Canva’s design features. It lets you make new designs, fill templates, find your designs, and save them as PDFs or images. Here’s a simple guide on how to install and get started with Canva MCP.

What You Need Before You Start

Before you install Canva MCP, you need a few things ready. First, you need an AI assistant that supports AI connectors or MCP servers, like ChatGPT, Claude, Gemini CLI, Cursor, or VS Code with the right extension. You also need a Canva account and any Canva plan. Some features like autofill might need a paid plan such as Canva Pro or Enterprise.

How to Connect Canva MCP to Your AI Assistant

Every AI assistant needs to be set up differently to work with Canva MCP. Below are steps for some popular assistants.

Connect with ChatGPT

To add Canva MCP to ChatGPT, start with these quick steps:

1. Open ChatGPT on your computer or web browser.

2. Start a new chat, click on Tools, then choose Use connectors.

3. Click Add sources and find the Canva connector. Enable it by following the screen instructions.

4. When asked, click Allow to give Canva MCP access to your Canva account.

This connects Canva MCP directly with ChatGPT.

Connect with Claude Using Connectors

Claude’s paid plan users can connect easily:

1. Open Claude on the web or desktop.

2. Start a new chat, click the Search and tools button, then pick Manage connectors.

3. Find the Canva connector and click Connect.

4. If you don’t see Canva, click Browse connectors, look under the Web tab, then add it.

5. Click Allow when Canva MCP asks for access to your Canva account.

Connect with Claude Using an MCP Server

You don’t need a paid plan here but will set up Canva MCP manually.

1. Open Claude Desktop.

2. Open settings using Ctrl + , on Windows or Command + , on Mac.

3. Go to the Developer tab.

4. Choose Edit Config and add this JSON configuration:

{
  "mcpServers": {
    "Canva": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@latest",
        "https://mcp.canva.com/mcp"
      ]
    }
  }
}

1. Save, close settings, and restart Claude Desktop.

2. When prompted, click Allow to give access to your Canva account.

Setting Up for Other Tools (Cursor, Gemini CLI, VS Code)

You can also set up Canva MCP with tools like Cursor, Gemini CLI, or VS Code.

• For Cursor, create a configuration file .cursor/mcp.json with the same JSON shown above.

• For VS Code, create .vscode/mcp.json with this slightly different format:

{
  "servers": {
    "Canva": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote@latest",
        "https://mcp.canva.com/mcp"
      ]
    }
  }
}

• For Gemini CLI, run this command in your project:

gemini extensions install https://github.com/canva-sdks/canva-gemini-extension

Restart Your AI Assistant and Approve Access

After configuration, save changes and restart your AI assistant app like Claude Desktop, Cursor, or VS Code. Your app will open a browser tab asking for permission to use Canva MCP. Click Allow to finish setup.

How to Test Canva MCP Connection

To make sure everything works, look for the Canva connector in your AI assistant interface. You can try this simple test command:

Show me my most recently edited Canva design

If connected right, the tool will ask for approval. Approve it and wait for the response.

---

This guide helps you quickly install and use Canva MCP with your AI assistant, enabling seamless design creation and management inside Canva.