Google Ads MCP Server

import type { CSSProperties, ReactNode } from 'react'; const headingStyle: CSSProperties = { fontSize: '2.5rem', fontWeight: 600, marginBottom: '0.5rem', }; const subheadingStyle: CSSProperties = { fontSize: '1.1rem', color: 'rgba(245, 247, 255, 0.75)', maxWidth: '46rem', lineHeight: 1.6, }; const sectionTitleStyle: CSSProperties = { fontSize: '1.5rem', fontWeight: 600, marginBottom: '0.75rem', }; const cardStyle: CSSProperties = { backgroundColor: 'rgba(15, 20, 32, 0.85)', border: '1px solid rgba(255, 255, 255, 0.08)', borderRadius: '16px', padding: '1.75rem', marginBottom: '1.5rem', boxShadow: '0 18px 60px rgba(6, 8, 14, 0.45)', }; const codeStyle: CSSProperties = { fontFamily: '"JetBrains Mono", "SFMono-Regular", Menlo, Monaco, Consolas, "Liberation Mono", "Courier New", monospace', fontSize: '0.95rem', background: 'rgba(12, 17, 28, 0.9)', borderRadius: '10px', padding: '1rem 1.25rem', display: 'block', overflowX: 'auto', margin: '0.75rem 0', border: '1px solid rgba(255, 255, 255, 0.04)', }; const listStyle: CSSProperties = { display: 'grid', gap: '1.1rem', paddingLeft: '1.25rem', margin: '0', }; const linkStyle: CSSProperties = { color: '#7dd3fc', textDecoration: 'underline', textDecorationColor: 'rgba(125, 211, 252, 0.4)', textDecorationThickness: '2px', }; const tableStyle: CSSProperties = { width: '100%', borderCollapse: 'collapse', }; const thStyle: CSSProperties = { textAlign: 'left', padding: '0.75rem 0.5rem', fontSize: '0.9rem', fontWeight: 600, borderBottom: '1px solid rgba(255, 255, 255, 0.1)', }; const tdStyle: CSSProperties = { padding: '0.75rem 0.5rem', fontSize: '0.95rem', verticalAlign: 'top', borderBottom: '1px solid rgba(255, 255, 255, 0.05)', lineHeight: 1.6, }; const accentStyle: CSSProperties = { display: 'inline-flex', alignItems: 'center', gap: '0.5rem', fontSize: '0.95rem', color: 'rgba(125, 211, 252, 0.85)', background: 'rgba(125, 211, 252, 0.12)', borderRadius: '999px', padding: '0.35rem 0.9rem', marginBottom: '1rem', }; const containerStyle: CSSProperties = { minHeight: '100vh', padding: '3.5rem 1.5rem 4.5rem', maxWidth: '960px', margin: '0 auto', fontFamily: '"Inter", system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif', }; const featureGridStyle: CSSProperties = { display: 'grid', gap: '1rem', marginTop: '1rem', }; const featureItemStyle: CSSProperties = { display: 'flex', gap: '0.75rem', }; const bulletStyle: CSSProperties = { fontSize: '1.5rem', lineHeight: 1, marginTop: '0.1rem', }; const rawBaseUrl = process.env.NEXTAUTH_URL?.trim(); const baseUrlSource = rawBaseUrl && rawBaseUrl.length > 0 ? rawBaseUrl : 'https://your-app.vercel.app'; const baseUrl = baseUrlSource.replace(/\/$/, ''); const appName = process.env.NEXT_PUBLIC_APP_NAME ?? 'Google Search MCP Server'; function InlineCode({ children }: { children: ReactNode }) { return ( <code style={{ fontFamily: codeStyle.fontFamily, background: 'rgba(12, 17, 28, 0.85)', borderRadius: '6px', padding: '0.15rem 0.45rem', fontSize: '0.92rem', border: '1px solid rgba(255, 255, 255, 0.05)', }} > {children} </code> ); } export default function HomePage() { return ( <main style={containerStyle}> <div style={accentStyle}> <span role="img" aria-hidden="true"> 🔍 </span> SEO research toolkit for MCP </div> <h1 style={headingStyle}>{appName}</h1> <p style={subheadingStyle}> Operate a Model Context Protocol server that taps directly into Google Suggest, Google Trends, and Search Console data. Batch keyword discovery, trend monitoring, and clustering workflows run inside ChatGPT or any MCP-compatible client—no custom middleware required. </p> <section style={cardStyle}> <h2 style={sectionTitleStyle}>Recent updates</h2> <div style={featureGridStyle}> <div style={featureItemStyle}> <span style={bulletStyle} aria-hidden="true"> • </span> <p> Autocomplete and Trends tools now accept single inputs (<InlineCode>query</InlineCode>{' '} / <InlineCode>keyword</InlineCode>) or batched arrays ( <InlineCode>queries</InlineCode> / <InlineCode>keywords</InlineCode>) for faster research loops. </p> </div> <div style={featureItemStyle}> <span style={bulletStyle} aria-hidden="true"> • </span> <p> Keyword clustering combines Google Search Console metrics with optional autocomplete expansions to recommend primary pages, supporting up to 1,500 terms per run. </p> </div> <div style={featureItemStyle}> <span style={bulletStyle} aria-hidden="true"> • </span> <p> MCP responses include consistent success and error envelopes ( <InlineCode>{'{ ok, data }'}</InlineCode> / <InlineCode>{'{ ok, error }'}</InlineCode> ) with matching HTTP and JSON-RPC behaviour. </p> </div> </div> </section> <section style={cardStyle}> <h2 style={sectionTitleStyle}>Quick start</h2> <ol style={listStyle}> <li> Install dependencies and run locally with <InlineCode>pnpm install</InlineCode> then{' '} <InlineCode>pnpm dev</InlineCode> (Node.js 18.18 or newer). </li> <li> Create a service account with access to your Search Console property and set{' '} <InlineCode>GOOGLE_CLIENT_EMAIL</InlineCode> and{' '} <InlineCode>GOOGLE_PRIVATE_KEY</InlineCode> (newline characters escaped as{' '} <InlineCode>\n</InlineCode>). Alternate variable names are respected for backwards compatibility. </li> <li> Add the MCP server endpoint <InlineCode>{baseUrl}/api/mcp</InlineCode> to ChatGPT Developer Mode or another MCP client and invoke the sample payloads below. </li> </ol> </section> <section style={cardStyle}> <h2 style={sectionTitleStyle}>Health check</h2> <p> Call <InlineCode>GET {baseUrl}/api/mcp</InlineCode> to confirm the server is online. A successful response looks like: </p> <pre style={codeStyle}> {`{ "status": "ok", "tools": ["ping", "get_autocomplete_suggestions", "get_trend_index", "get_keyword_clusters"] }`} </pre> </section> <section style={cardStyle}> <h2 style={sectionTitleStyle}>Tool invocations</h2> <table style={tableStyle}> <thead> <tr> <th style={thStyle}>Tool</th> <th style={thStyle}>Sample payload</th> </tr> </thead> <tbody> <tr> <td style={tdStyle}> <strong>Ping</strong> </td> <td style={tdStyle}> <pre style={codeStyle}> {`{ "tool": "ping" }`} </pre> </td> </tr> <tr> <td style={tdStyle}> <strong>Autocomplete suggestions</strong> </td> <td style={tdStyle}> <pre style={codeStyle}> {`{ "tool": "get_autocomplete_suggestions", "input": { "queries": ["toroidal transformer", "toroidal core"] } }`} </pre> </td> </tr> <tr> <td style={tdStyle}> <strong>Trend index</strong> </td> <td style={tdStyle}> <pre style={codeStyle}> {`{ "tool": "get_trend_index", "input": { "keywords": ["toroidal transformer", "toroidal core"], "timeRange": "today 12-m", "geo": "US" } }`} </pre> </td> </tr> <tr> <td style={tdStyle}> <strong>Keyword clusters</strong> </td> <td style={tdStyle}> <pre style={codeStyle}> {`{ "tool": "get_keyword_clusters", "input": { "query": "solar panels", "siteUrl": "https://example.com/", "timeRange": "last_90_days", "maxKeywords": 1500 } }`} </pre> </td> </tr> </tbody> </table> </section> <section style={{ ...cardStyle, marginBottom: 0 }}> <h2 style={sectionTitleStyle}>Need help?</h2> <p> Review the{' '} <a style={linkStyle} href="https://github.com/openai/google-search-mcp"> project README </a>{' '} for deployment tips, troubleshooting guidance, and changelog-style updates. </p> </section> </main> ); }