If you build AI agents in Cursor or Claude Desktop, you can now work with Kynver — listing your agent, checking verification status, even starting the verification process — without ever leaving your coding environment.

This guide explains what that means and how to set it up in under five minutes.

What is the Kynver MCP server?

MCP stands for Model Context Protocol — it's a standard that lets AI coding tools (like Cursor and Claude Desktop) connect to external services as tools that the AI can use on your behalf.

The Kynver MCP server gives your AI assistant a set of Kynver-specific tools it can call while you work:

In plain English: your AI assistant becomes a hands-free way to manage your Kynver presence while you're building.

What you'll need before you start

Getting your API key (optional, for write tools)

If you only want to look up agents and check verification status, skip this step — no key needed. If you want your AI assistant to register agents or advance verification on your behalf, you'll need a key.

  1. Sign in at kynver.com/dashboard.
  2. Click your agent in the sidebar.
  3. Go to the Settings tab.
  4. Find API Keys and click Create key.
  5. Copy the key and keep it somewhere safe — you won't be able to see it again.

Setting up in Cursor

Cursor has built-in MCP support. Here's how to add the Kynver server:

  1. Open Cursor and go to Settings (the gear icon, or Cmd+, on Mac / Ctrl+, on Windows).
  2. In the left sidebar, find MCP and click it.
  3. Click Add new MCP server.
  4. Fill in the fields:
    • Name: kynver
    • Command: npx
    • Arguments: -y @kynver-app/mcp-server
  5. If you have an API key, add an environment variable:
    • Key: KYNVER_API_KEY   Value: your key
  6. Click Save. Cursor will download and start the Kynver server automatically.

Alternatively, if you prefer editing config files directly, open (or create) .cursor/mcp.json at the root of your project and add:

{
  "mcpServers": {
    "kynver": {
      "command": "npx",
      "args": ["-y", "@kynver-app/mcp-server"],
      "env": {
        "KYNVER_API_KEY": "<your-api-key>"
      }
    }
  }
}

Leave out the KYNVER_API_KEY line entirely if you don't have one yet — read-only tools will still work.

Setting up in Claude Desktop

  1. Open Claude Desktop and go to Settings.
  2. Click the Developer tab.
  3. Click Edit Config. This opens a config file in a text editor.
  4. Add the Kynver section inside "mcpServers":
{
  "mcpServers": {
    "kynver": {
      "command": "npx",
      "args": ["-y", "@kynver-app/mcp-server"],
      "env": {
        "KYNVER_API_KEY": "<your-api-key>"
      }
    }
  }
}
  1. Save the file.
  2. Restart Claude Desktop completely (quit and reopen). It needs a fresh start to pick up the new server.

On macOS, the config file lives at:
~/Library/Application Support/Claude/claude_desktop_config.json

On Windows, it lives at:
%APPDATA%\Claude\claude_desktop_config.json

Setting up in VS Code (GitHub Copilot)

  1. In the root of your project, create a file called .vscode/mcp.json.
  2. Add this content:
{
  "servers": {
    "kynver": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@kynver-app/mcp-server"],
      "env": {
        "KYNVER_API_KEY": "<your-api-key>"
      }
    }
  }
}
  1. Save the file. VS Code will detect the server and connect automatically.

How to use it once it's set up

Once connected, just ask your AI assistant naturally — it knows what tools are available and will use them when relevant. For example:

The AI will call Kynver on your behalf and summarise the result right in the chat.

Write tools (register, verify ownership, start KYC, etc.) require a KYNVER_API_KEY. If you get a "401 Unauthorized" or "API key required" message, add your key to the config as described above.

What does "npx" actually do?

npx is a tool that comes with Node.js. When your AI IDE runs npx -y @kynver-app/mcp-server, it automatically downloads the latest version of the Kynver MCP server from the npm registry and runs it. You don't need to install anything separately — npx handles it. The -y flag just means "yes, go ahead" so it doesn't pause to ask.

Troubleshooting

Ready to connect Kynver to your AI IDE? The full package is on npm — no install needed, just npx -y @kynver-app/mcp-server. Questions? Reach us at hello@kynver.com.

Go to your dashboard →