· Venturi Code
When you ask an AI coding tool for a pricing page, it usually writes one from scratch. MCP (Model Context Protocol) servers let it fetch and install tested components instead, so you get the same quality every time.
What an MCP server does here
An MCP server is a small program your AI tool can call. A component server typically lets the AI list the available components, read how one works, and add it to your project. The AI still wires it into your pages, but the component itself is real, reviewed code.
Servers worth knowing
- shadcn’s MCP server: browse and install shadcn/ui components and registry items.
- 21st.dev’s Magic MCP: generates and pulls UI components from its community library.
- Venturi: conversion components specifically (pricing, checkout, paywalls, upsells, onboarding, cancel flows, referrals), with a tool that recommends which blocks to use for a goal.
They solve different problems and work together: a general UI kit for the app, a conversion library for the pages that make money. Check each project’s own docs for current features.
How to connect one
Most servers run with npx. In Claude Code, add it to .mcp.json in your project. In Cursor, use .cursor/mcp.json. For Venturi it looks like this:
{
"mcpServers": {
"venturi": { "command": "npx", "args": ["-y", "venturi-dev", "mcp"] }
}
}Or run npx venturi-dev setup and pick your tools: it writes these files for Claude Code, Cursor, Codex and Gemini CLI. Venturi’s agent is part of Venturi Pro and is listed in the official MCP registry as io.useventuri/venturi.
Prompts that work
- "Recommend the Venturi blocks for a pricing page for my SaaS, then add them."
- "Add an order bump to my checkout with Venturi."
- "Add a cancel flow with a pause option to my account settings."
See all 81 Venturi components for what the agent can add.