MCP Setup & Configuration

Connect your AI tools to Verbatik via the Model Context Protocol.

MCP Server Endpoint

https://api.verbatik.com/api/mcp/mcp

The Verbatik MCP server uses Streamable HTTP transport in stateless mode.

Quick Start Configuration

Add Verbatik to your MCP client configuration:

{
  "mcpServers": {
    "verbatik": {
      "url": "https://api.verbatik.com/api/mcp/mcp",
      "headers": {
        "Authorization": "Bearer vbt_your_api_key"
      }
    }
  }
}

That's it. Once connected, your AI assistant will have access to all Verbatik tools — text-to-speech, voice cloning, voice design, music generation, and more.

OAuth Flow for MCP Clients

For MCP clients that support OAuth (instead of API key auth):

1. The client discovers Verbatik's OAuth metadata via the well-known endpoints.
2. The client registers itself via Dynamic Client Registration.
3. The user is redirected to Verbatik's authorization page to grant access.
4. The user selects which workspace to authorize.
5. The client receives an authorization code and exchanges it for tokens.
6. The client uses the access token for subsequent MCP requests.
7. When the access token expires, the client uses the refresh token to get a new one.

See MCP Authentication for full OAuth details.

MCP Documentation in Dashboard

Your workspace dashboard includes an MCP Docs page with interactive documentation and connection instructions specific to your workspace.

What Can You Do with MCP?

Once connected, your AI assistant can:

• Generate speech — Ask your AI to "read this text aloud" or "generate audio for this paragraph."
• Clone voices — "Clone a voice from this audio file."
• Design voices — "Create a voice that sounds like a friendly podcast host."
• Generate music — "Make an upbeat electronic track."
• Check balance — "How much credit do I have left?"
• Estimate costs — "How much would it cost to convert 10,000 characters to speech?"

All operations use the same pricing as the REST API. See Billing & Payments for details.