Vomyra MCP Server

What is Vomyra MCP Server?

Vomyra runs a hosted MCP (Model Context Protocol) server that any compatible AI tool can connect to. Once connected, you can manage your Vomyra voice assistants, tools, phone numbers, and calls through natural language � no clicking through the dashboard required.

Think of it as giving Claude or Cursor a direct line into your Vomyra account. You describe what you want in plain English, and the AI figures out which Vomyra actions to take and runs them for you.

Why use it?

If you use AI tools in your daily work, or want to automate repetitive Vomyra tasks without touching the dashboard, MCP Server is the fastest way to get there.

How it works

It's a one-time setup. After that, every conversation with your AI tool has live access to your Vomyra account.

Connect from Claude.aiRecommended

Claude.ai handles the sign-in flow automatically. You only need to add the connector once � it stays connected and refreshes quietly in the background.

1. 1

   Open Connectors settings

   Go to claude.ai ? Settings ? Connectors ? Add custom connector.

   

2. 2

   Enter the MCP URL

   Give it a name (e.g. Vomyra) and paste:https://vomyra.com/mcp

   

3. 3

   Sign in with your Vomyra account

   Choose OAuth when prompted. Sign in through the popup and approve access � the mcp:tools permission gives Claude access to your assistants, tools, calls, and numbers. You'll be redirected back automatically.

   

4. 4

   Start using Vomyra tools

   The connector becomes active in your conversations. Verify with: "List my Vomyra assistants."

   

Example � Claude listing your Vomyra assistants via MCP

Connect from Claude Desktop

Add Vomyra to claude_desktop_config.json:

• macOS:~/Library/Application Support/Claude/
• Windows:%APPDATA%\Claude\

{
  "mcpServers": {
    "vomyra": {
      "url": "https://vomyra.com/mcp",
      "type": "http"
    }
  }
}

Connect from Cursor / Windsurf

Use an API key (see API key setup below) for the simplest setup:

{
  "mcpServers": {
    "vomyra": {
      "url": "https://vomyra.com/mcp",
      "type": "http",
      "headers": {
        "x-api-key": "<your Vomyra API key>"
      }
    }
  }
}

Custom code (TypeScript)

import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StreamableHTTPClientTransport } from "@modelcontextprotocol/sdk/client/streamableHttp.js";

const transport = new StreamableHTTPClientTransport(
  new URL("https://vomyra.com/mcp"),
  {
    requestInit: {
      headers: { "x-api-key": process.env.VOMYRA_API_KEY! },
    },
  }
);

const client = new Client({ name: "my-app", version: "1.0.0" });
await client.connect(transport);

const { tools } = await client.listTools();
console.log(tools.map((t) => t.name));
// ? ["list_assistants", "create_assistant", "initiate_call", ...]

Authentication

There are two ways to authenticate � one for interactive use, one for code and automation:

API Key

Recommended for backends

Generate a key from Dashboard ? API Keys, then send it on every request:

HTTP header (either works)

x-api-key: <your key>

# or

Authorization: Bearer <your key>

When to use Vomyra MCP Server

MCP Server works well for most assistant management and automation scenarios today. A few areas are still being developed � check the limitations section before building around them.

Example workflows

These are real prompts you can try once Vomyra is connected to your AI tool.

create_assistant(name: "Aarti", system_prompt: "...", welcome_message: "...")

list_assistant_options(section: "voice")
update_assistant_voice(assistant_id: ..., voice_provider: "azure", voice_name: "hi-IN-AartiNeural")
update_assistant_model(assistant_id: ..., ai_provider: "groq", model: "llama-3.3-70b-versatile")

initiate_call(assistant_id: ..., customer_number: "+919876543210", customer_name: "Pankaj", call_mode: "demo")

list_tool_types()  ? returns catalog with _id values

create_tool(
  tool_catalog_id: "<apiRequest id>",
  name: "lookup_order",
  description: "Looks up a Spice Garden order by id",
  function_config: {
    request_url: "https://api.spicegarden.com/v1/orders",
    request_http_method: "GET",
    authentication: { type: "bearerToken", ... },
    parameters: [{ name: "order_id", type: "string", required: true }]
  }
)

list_assistants()
assign_tools_to_assistant(assistant_id: ..., tool_ids: [...])

list_calls(limit: 5)
get_call(call_id: ...)  ? for each failed call

Rate limits

Every response includes headers showing how many requests you have left (X-RateLimit-Remaining) and your limit (X-RateLimit-Limit). If you hit the limit, you'll get a 429 error � wait until the time shown in Retry-After before trying again.

Error model

Connection-level problems (like authentication failures or rate limits) come back as error objects. Errors within a specific tool (like 'assistant not found') come back as a normal tool result marked with isError: true � not as connection errors.

Tools at a glance

All 24 tools and 3 read-only resources are scoped to your account � everything you do through MCP affects only your own assistants, tools, numbers, and calls.

These are read-only snapshots your AI tool can use as context. Think of them as quick lookup pages � ask for the resource and get back the current data for your account.

Need any of these via MCP? File a request with support@vomyra.com.