Context7
Feeds AI the latest docs for libraries and frameworks before it answers, so code it writes does not use outdated, removed functions.
Description from the maker (original English)
Google Search Console, GA4 and PageSpeed MCP server for Claude, ChatGPT and Cursor. Free, hosted.
These three are never merged into one score: opinions, measurements and popularity are different things.
We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.
Pick the AI app you use, then copy the settings.
You need
Which app do you use?
Pick your app above. We remember it, so next time the button says “Install in …”.
Works in Claude on the web, desktop and mobile:
https://mcp.gscpap.com/mcpRun this in your terminal
claude mcp add --transport http gsc-pap https://mcp.gscpap.com/mcpPaste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json
{
"mcpServers": {
"gsc-pap": {
"url": "https://mcp.gscpap.com/mcp"
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
Paste into .vscode/mcp.json in your project
{
"servers": {
"gsc-pap": {
"type": "http",
"url": "https://mcp.gscpap.com/mcp"
}
}
}Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.
ChatGPT connects only to servers on the internet:
https://mcp.gscpap.com/mcpMenu names can differ slightly by plan and over time.
Your choice stays on this device in a cookie; no account needed.
These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.
The server asks you to sign in before it lists its tools.
We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 8, 2026
The registry name was verified against the domain gscpap.com (the registry checks domain ownership)
We have not read this repo from GitHub yet
We have not read this repo from GitHub yet
Responded in 1035 ms (HTTP 401)
Declares OAuth sign-in, so your AI app can take you to log in
Not checked yet; we read the tool list when we check the endpoint
Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.
It passed 3 of 3 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.
The registry does not list a required key, though some servers ask you to sign in on first use.
It runs in the maker’s cloud. Connect over the internet with nothing to install, but what you send passes through the maker’s servers.
The GitHub repo last changed –, and the latest version (0.3.5) reached the registry 2 days ago.
Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. It connects by URL, so apps that accept remote MCP servers work too, such as ChatGPT in developer mode.
In the same category: Context7, Codebase Memory, Exa
Servers in the same category from other makers, sorted by stars.
Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi
Feeds AI the latest docs for libraries and frameworks before it answers, so code it writes does not use outdated, removed functions.
Codebase knowledge graph for AI agents — 162 languages, sub-ms queries, 99% fewer tokens.
Connect AI agents to Exa for web search, content fetching, and multi-step research.
Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.
Ask Claude about your Google Search Console data: top queries, decaying pages, ranking trends, quick wins, cannibalization, Core Web Vitals, and GA4 conversions per landing page.
Free. 38 read-only tools. Hosted, so there is no Google Cloud project, no Python and no JSON credentials to manage. You sign in with Google at gscpap.com and connect; the OAuth client is already registered, for everyone.
| Price | Free, $0. No credit card, no trial clock, no paid tier today. Pricing |
| Tools | 38, all read-only |
| Data sources | Search Console, Google Analytics 4, PageSpeed Insights, live Google SERPs |
| MCP endpoint | https://mcp.gscpap.com/mcp (streamable HTTP) |
| Auth | Google sign-in. OAuth 2.0 with Dynamic Client Registration for connectors, or a bearer token for config-file clients |
| Scopes | webmasters.readonly, plus analytics.readonly if you connect GA4. It cannot change a property, submit a sitemap or remove a URL |
| History | A Personal Archive copies the 16 months Google still has when you connect, then keeps every new day, so your history grows past Google's 16-month limit |
| Registry | com.gscpap/gsc-pap in the official MCP Registry |
ga4_ai_traffic: the visits ChatGPT, Perplexity, Gemini, Claude and Copilot send you, per engine, against Google organic, and which pages they land on. It also counts the assistant visits GA4 files under Referral instead of its AI Assistant channelfind_ai_overview_suspects: queries whose CTR fell while their position held, the mark an AI Overview leaves in Search ConsoleEvery change, with dates: gscpap.com/changelog
A single MCP server (gsc-pap) exposing 38 tools over the Search Console, Google Analytics 4, PageSpeed Insights and SERP APIs:
Spot quick SEO wins
find_quick_wins: queries at positions 5 to 20 with at least 100 impressions but under 5% CTR, which is where a title or snippet rewrite paysfind_cannibalization: several of your URLs competing for one query, splitting the clicksFind what is losing traffic
find_content_decay: pages losing organic traffic over timehistorical_position_trend: one query's position over the last N days, trending up, down or just volatilecompare_periods: diff any two windows, for example the last 28 days against the prior 28find_dead_pages: pages appearing for months that have never earned a click, split by whether Google shows them and nobody clicks or barely shows them at allBeat the 16-month limit
search_analytics_archive: query your full history, past Google's 16-month retentionarchive_status: per-site coverage, earliest and latest day, row count, last runSee the SERP your rankings live in
serp_top_results: the live top organic results for a query and country, plus which SERP features are present (featured snippet, map pack, People Also Ask; AI Overviews are reported as unknown)serp_competitor_analysis: where your domain sits for a query and the full top 10 around itserp_features_audit: up to 25 queries at once, to explain a low CTR at a good positionDiagnose performance and Core Web Vitals
pagespeed_insights: Lighthouse audit, the four scores, Core Web Vitals and the top opportunitiespagespeed_field_data: real-user CrUX metrics (LCP, INP, CLS at p75) for a URL and for the origin, in about a second, with an optional weekly trend of the last 25 weeksDiagnose indexing issues
get_index_coverage_summary: account-wide breakdown of indexing statesinspect_url: one URL in depth, straight from the URL Inspection APIbatch_inspect_urls: bulk URL inspection, up to 2,000 URLsfind_new_pages: pages Search Console started showing this week, and whether each has earned a click yetPull performance data
get_performance_overview: clicks, impressions, CTR and position in one line, for any date range or the last N dayssearch_analytics_query: the whole Search Analytics surface, any dimension and filterget_search_by_page_query: the query mix behind one pageSee how AI assistants send you traffic (GEO)
ga4_ai_traffic: visits from ChatGPT, Perplexity, Gemini, Claude and Copilot, per engine, against organic, and the pages they send people to; property_id="all" compares every sitefind_ai_overview_suspects: queries whose CTR fell while their position held, the footprint an AI Overview leavesConnect search to engagement and revenue (GA4)
Every GA4 report tool takes days, or an exact start_date and end_date.
list_ga4_properties: find the property_id the other GA4 tools needga4_overview: users, sessions, views, bounce and engagement, with a daily breakdownga4_top_pages: top landing pages by sessionsga4_traffic_sources: channel, source, medium, campaignga4_realtime: active users in the last 30 minutesga4_conversions_by_source: source and medium with conversions and revenuega4_user_demographics: country, device and browser slicesga4_engagement: engagement rate, duration and bounce per pagega4_events_by_name: scroll depth, outbound clicks, downloads and custom events, optionally per pagega4_compare_periods: diff any window for any GA4 metric, including an exact start and end datecorrelate_gsc_to_ga4: the cross-tool one. Search Console clicks per page joined to GA4 sessions, conversions and revenue for that same pageManage sites and sitemaps
list_sites, get_site_details: your verified propertieslist_sitemaps, get_sitemap_details: sitemap healthdetect_duplicate_properties: domain and URL-prefix properties that split your reporting in twoUse it in chat:
Show me my 5 worst-decaying pages last 30 days, with the queries they used to rank for.
Which pages are stuck at position 11 to 20 with 1k+ impressions? Suggest content tweaks for the top 3.
Run a Lighthouse audit on https://example.com/pricing and list the top 3 opportunities by potential ms savings.
Pull my Search Console data from January 2024, which is past Google's 16-month retention.
Correlate my Search Console pages with GA4 conversions for the last 28 days. Which top-clicked pages bring zero conversions?
Why is my CTR 1% at position 4 for this query? Check what else is on that SERP.
https://mcp.gscpap.com/mcpThe server implements OAuth 2.0 with Dynamic Client Registration (RFC 7591), so the client registers itself. Step by step, with screenshots: Add a Search Console connector to Claude.ai.
ChatGPT supports the same flow, but custom MCP connections there are a Business and Enterprise/Edu feature on ChatGPT web, which is OpenAI's rule rather than ours: Connect ChatGPT to Google Search Console.
If you would rather drive the CLI yourself:
claude mcp add --transport http --scope user gsc-pap https://mcp.gscpap.com/mcp \
--header "Authorization: Bearer YOUR_TOKEN_HERE"
This plugin variant reads a ${GSC_PAP_TOKEN} env var instead. Set it once:
# macOS / Linux
echo 'export GSC_PAP_TOKEN="paste-your-token-here"' >> ~/.zshrc source ~/.zshrc
Restart Claude Code, or run claude mcp list to check, and the gsc-pap server shows up as Connected.
Create a token at app.gscpap.com/account/tokens, then add this to ~/.cursor/mcp.json (Windows: %USERPROFILE%\.cursor\mcp.json):
{
"mcpServers": { "gsc-pap": { "url": "https://mcp.gscpap.com/mcp", "headers": { "Authorization": "Bearer YOUR_TOKEN" } } } }
Full guide: Connect Cursor to Google Search Console.
Add this to .vscode/mcp.json. VS Code asks for the token once and stores it as a secret:
{
"servers": { "gsc-pap": { "type": "http", "url": "https://mcp.gscpap.com/mcp", "headers": { "Authorization": "Bearer ${input:gsc_pap_token}" } } }, "inputs": [ { "type": "promptString", "id": "gsc_pap_token", "description": "GSC PAP bearer token (from app.gscpap.com/account/tokens)", "password": true } ] }
Full guide: Connect GitHub Copilot in VS Code to Google Search Console.
Claude Desktop reaches remote servers through mcp-remote, so it needs Node.js. Add this to ~/Library/Application Support/Claude/claude_desktop_config.json (Windows: %APPDATA%\Claude\claude_desktop_config.json) and restart Claude Desktop:
{
"mcpServers": { "gsc-pap": { "command": "npx", "args": ["-y", "mcp-remote", "https://mcp.gscpap.com/mcp", "--header", "Authorization:${AUTH}"], "env": { "AUTH": "Bearer YOUR_TOKEN" } } } }
Full guide: Connect Claude Desktop to Google Search Console. You can also skip the file and add it as a connector, as in the Claude.ai section above.
Setup guides for the other clients (Cursor, VS Code with Copilot, Cline, opencode, Codex CLI, Gemini CLI, Windsurf, Zed, JetBrains, Continue, Warp, Goose, LM Studio, ChatGPT, plain HTTP) are at gscpap.com/integrations.
Bearer and a space in front where the config asks for the whole header.gsc-pap sits inside mcpServers (servers in VS Code), then restart the client.Only the beginning is shown; read the rest in the repo