The local MCP server runs on your own machine and communicates over stdio transport. It suits desktop AI applications and gives you the most control over execution.
Local or remote?
Local MCP | Remote MCP | |
Where it runs | Your machine | Cloud-hosted by Tomba |
Transport | Stdio | HTTP / SSE |
Installation | Required | None |
Best for | Desktop AI apps | Web applications |
Before you start
You will need your Tomba API key and secret. See How to find my API key and Where can I find my secret key.
You also need Node.js available, since the server runs through npx.
Configuration
{ "mcpServers": { "tomba": { "command": "npx", "args": ["-y", "@tomba-mcp-server"], "env": { "TOMBA_API_KEY": "your-api-key", "TOMBA_API_SECRET": "your-api-secret" } } }}
The server starts automatically when your AI tool launches — there is no separate process to manage.
What it provides
Email discovery — Email Finder by first name, last name and domain; Domain Search across a company; LinkedIn Finder from a profile URL; Author Finder from an article URL.
Verification — Email Verifier for deliverability, plus status checking.
Security
Your credentials stay in the local configuration file on your machine. They are never transmitted to AI model providers.
Related
Related Resources
API Documentation — Complete API reference and guides
Authentication Guide — How to authenticate API requests
Official Libraries — SDKs for Python, Node.js, PHP, Go, and more
Rate Limits — API rate limiting and best practices
Domain Search — Find all emails for a company domain
Domain Search API — API reference for domain search
Email Finder — Find professional email addresses
Email Finder API — API reference for email finder