
Hey everyone! We’re Wen and Tobias and we build Metorial -- the enterprise-grade infra layer for MCP.
Magic MCP let’s you connect any MCP client, like Cursor, Claude Code, or Claude Desktop, to more than 600 integrations ranging from Sentry or Linear all the way to Salesforce or QuickBooks.
Metorial provides infra for deploying, connecting to, and securing MCP servers for developers and enterprises. We are working with startups, unicorns, and fortune 500 companies to let them use MCP reliably and effectively. Metorial let’s them get AI agents into production in hours, not weeks.
Yet, many of our customers still struggle with the basics, like integrating MCP servers into Cursor and Claude Code. Today, we’re solving that.
Magic MCP let’s you connect vibe coding tools to 600+ integrations (you can even create your own MCP servers and deploy them to Metorial) -- zero configuration required.
Magic MCP uses the same infra that already powers thousands of MCP server deployments. You get enterprise-grade security, scaling, and reliability, sub-second boot up times, and full observability.
We handle all of the complicated parts of MCP so our customers get to enjoy the good parts. Enterprises love us because we solve security, scalability, and access control for MCP. Developers and startups love us because they can deploy MCP servers on our platform in just 3 clicks -- we even handle OAuth. Our SDKs let them connect their agents to MCP servers on Metorial in a single function call. Our serverless runtime is the only one that solves security for multi-tenancy, supports sub-second cold start, and does stateful hibernation for long-running connections.
Struggling with MCP or building agents that need integrations?
Try metorial.com. We can help you integrate Metorial or Magic MCP and add any MCP servers you need.
Book a call: cal.com/karim-rahme/launch-yc
Email us: founders@metorial.com
Tobias and Wen have been friends for over 10 years. They met in a top technical high school in Austria where they worked on many projects together. Later, they co-lead the technical development at Valicit, a ticketing startup based in Abu Dhabi.
When MCP was released, we quickly understood its value in offering a natural language interface for LLMs to execute external tools autonomously.
But when we tried to take MCP beyond client-side apps like Cursor, setting up the infra for security and scaling turned out to be a massive time sink.
Thatʼs when it clicked: developers like us need a turn-key solution to plug MCP into their agents.
So we pivoted from hacking infra together to building Metorial: the Vercel for MCP.
Integrations transform LLMs from simple chatbots into agents that can perform real tasks.
Recently, MCP has become the standard for connecting AI agents to external tools.
MCP works great locally, but production deployments are painful.
Running MCP servers in production means managing Docker configs, handling OAuth flows, and debugging issues without proper observability.
Developers want to ship agents now, not build infrastructure for months.