import {
DocsTitle,
DocsLead,
DocsH2,
DocsH3,
DocsP,
DocsList,
DocsLi,
DocsCode,
Mono,
} from '@/components/docs-page';
export const metadata = { title: 'Quickstart — BuildMyMCPServer docs' };
export default function Quickstart() {
return (
<>
Quickstart
Describe the tool you want, paste in any credentials, watch the build stream, copy a snippet
into your AI client. Five minutes from first prompt to a live OAuth-protected MCP server.
Prerequisites
An AI client that speaks MCP — Claude Desktop, Cursor, ChatGPT Custom Connectors, VS Code Copilot, or Continue.dev.
API credentials for whatever you want your server to access (Notion, your DB, etc.). Or pick the echo example to skip this.
1. Sign in
Hit the dashboard and enter your email. We send a magic link — no password.
2. Describe your tool
Click + New server and write what you want in plain language. The clearer
you are about which APIs and which tool names, the better the spec.
3. Confirm the plan
Step 2 of the wizard shows you exactly which tools Claude parsed from your prompt, the
input schemas, and which credentials we need from you. Fill them in. Skip the step
entirely for self-contained demo servers like the echo template.
4. Watch the build stream
The build goes through five states live over WebSocket:
queued → spec validated, job in BullMQ
generating → Claude returns spec (or cached preview is reused)
building → TypeScript rendered, static checks, Docker image built
deploying → container booted on an allocated host port
live → endpoint responds, OAuth gate is active
5. Install in your client
The Done screen shows three tabs — Claude Desktop, Cursor, ChatGPT — each with a copy-ready
snippet. Paste the JSON into your client's MCP config and restart. The OAuth handshake
runs automatically on first tool call.
What's next
Read about the underlying MCP concepts,
learn how the OAuth 2.1 flow protects each server,
or jump to authoring custom tools.
>
);
}