Google Search Console MCP
# <img src="./assets/a1-logo.svg" alt="A1" width="40"> Google Search Console MCP
**English** | [Русский](./README.ru.md)
[](https://www.npmjs.com/package/mcp-google-search-console)
[](https://glama.ai/mcp/servers/A1-x-Tech/mcp-google-search-console)
[](https://github.com/A1-x-Tech/mcp-google-search-console/actions/workflows/ci.yml)
[](./LICENSE)
**A1 Google Search Console MCP** connects an AI app to Google Search Console. Investigate search performance, check whether a URL is indexed, inspect sitemaps and deliberately submit or remove a sitemap when needed.
It works with the properties your Google account can access. The important detail is that it uses the exact Search Console property value — a domain property and a URL-prefix property are different objects.
- **18 tools.** Seven tools read properties, search data, sitemaps and index status; two add a property or submit a sitemap; three can remove data or call an arbitrary API method.
- **Connects from the conversation.** Say "connect Google Search Console": the server walks you through the OAuth client, catches Google's redirect on `127.0.0.1` with PKCE and keeps the tokens itself — no config files, no restart.
- **Exact property IDs.** `https://example.com/`, `https://www.example.com/` and `sc-domain:example.com` are distinct. `list_sites` shows the value to use.
- **Search data with context.** Query clicks, impressions, CTR and position by date, page, query, country, device or search appearance.
- **Indexing, not publishing.** URL inspection explains Google’s current status; it does not force a page into the index.
Start with a read-only question:
> Show the top 20 search queries for my property over the last 28 days, with clicks and CTR.
[Connect the server](#quick-start) · [Explore use cases](#what-you-can-ask-it-to-do) · [Open technical documentation](#technical-documentation)
---
## See it work in a minute
> **You:** Is `https://example.com/pricing` indexed? If not, why?
>
> **Assistant:** Inspects the URL and shows the index verdict, coverage, crawl information and canonical URLs. Nothing changes.
>
> **You:** Check my submitted sitemaps and prepare a resubmission for the one with errors.
>
> **Assistant:** Shows the sitemap, its warnings and errors, then asks for confirmation before submitting it again.
>
> **You:** Confirm.
>
> **Assistant:** Resubmits the selected sitemap. It does not change page content or guarantee indexing.
## Contents
- [Quick start](#quick-start)
- [What you can ask it to do](#what-you-can-ask-it-to-do)
- [How Search Console properties work](#how-search-console-properties-work)
- [What can change](#what-can-change)
- [Getting access](#getting-access)
- [Configuration](#configuration)
- [Data, limits and background work](#data-limits-and-background-work)
- [Technical documentation](#technical-documentation)
- [Support](#support)
## Quick start
You need Node.js 20+ and a Google account. Credentials are not required at install time — the server connects from the conversation.
1. Add the server to your AI app.
2. Say "connect Google Search Console": the assistant walks you through [creating the OAuth client and approving access](#getting-access) without editing config files.
3. Start with the read-only question above.
<details open><summary><strong>Codex</strong></summary>
<br>
<<<<<<< Updated upstream
**In the app:** open **Settings → MCP servers**, select **Add server**, choose **STDIO**, enter the command `npx -y mcp-google-search-console@latest` and environment variables `GOOGLE_SEARCH_CONSOLE_CLIENT_ID`, `GOOGLE_SEARCH_CONSOLE_CLIENT_SECRET`, `GOOGLE_SEARCH_CONSOLE_REFRESH_TOKEN`, then select **Save** and **Restart**.
||||||| Stash base
**In the app:** open **Settings → Plugins → MCP servers**, choose **Add server**, then add `npx -y mcp-google-search-console@latest` with `GOOGLE_SEARCH_CONSOLE_CLIENT_ID`, `GOOGLE_SEARCH_CONSOLE_CLIENT_SECRET` and `GOOGLE_SEARCH_CONSOLE_REFRESH_TOKEN`.
=======
**In the app:** open **Settings → Plugins → MCP servers**, choose **Add server**, then add `npx -y mcp-google-search-console@latest`.
>>>>>>> Stashed changes
```bash
codex mcp add google-search-console \
-- npx -y mcp-google-search-console@latest
codex mcp list
```
[Codex MCP documentation](https://learn.chatgpt.com/docs/extend/mcp?surface=cli)
</details>
<details><summary><strong>Claude Code</strong></summary>
<br>
```bash
claude mcp add \
--transport stdio --scope user google-search-console \
-- npx -y mcp-google-search-console@latest
claude mcp list
```
[Claude Code MCP documentation](https://code.claude.com/docs/en/mcp)
</details>
<details><summary><strong>Claude Desktop</strong></summary>
<br>
The current official path is **Settings → Extensions**. For a custom desktop extension, open **Advanced settings → Extension Developer → Install Extension…**, select a `.mcpb` file and follow the prompts.
This repository currently publishes an npm stdio package and does not contain a `.mcpb` bundle. For Claude Desktop builds that still support local configuration, use the following JSON stdio configuration as a fallback:
```json
{"mcpServers":{"google-search-console":{"command":"npx","args":["-y","mcp-google-search-console@latest"]}}}
```
In those builds, save it to `~/Library/Application Support/Claude/claude_desktop_config.json` on macOS or `%APPDATA%\Claude\claude_desktop_config.json` on Windows.
[Claude Desktop MCP documentation](https://support.claude.com/en/articles/10949351-getting-started-with-local-mcp-servers-on-claude-desktop)
</details>
<details><summary><strong>Cursor</strong></summary>
<br>
Add to `~/.cursor/mcp.json` on macOS/Linux or `%USERPROFILE%\.cursor\mcp.json` on Windows:
```json
{"mcpServers":{"google-search-console":{"type":"stdio","command":"npx","args":["-y","mcp-google-search-console@latest"]}}}
```
[Cursor MCP documentation](https://cursor.com/docs/mcp)
</details>
<details><summary><strong>VS Code</strong></summary>
<br>
Run **MCP: Open User Configuration** and add:
```json
{"servers":{"google-search-console":{"type":"stdio","command":"npx","args":["-y","mcp-google-search-console@latest"]}}}
```
Check it with **MCP: List Servers**. [VS Code MCP documentation](https://code.visualstudio.com/docs/agent-customization/mcp-servers)
</details>
## What you can ask it to do
### Find search opportunities
- Which queries and pages brought the most clicks this month?
- Which pages lost clicks compared with the previous period?
- Show queries containing `mcp` where average position is below 10.
### Check index status and sitemaps
- Is this URL indexed? Show coverage, crawl and canonical information.
- Which submitted sitemaps have errors or warnings?
- Resubmit this sitemap after showing its current status.
### Manage properties carefully
- List the Search Console properties I can access.
- Add this exact property value; I will verify ownership separately.
- Remove this property from my account after confirmation.
## How Search Console properties work
A URL-prefix property must include its protocol and trailing slash, for example `https://example.com/`. A domain property is written as `sc-domain:example.com`. A near-match causes 403 or 404, so use the exact value returned by `list_sites`.
`add_site` only registers a property. Verification remains in the Search Console UI or Site Verification API. Search data uses Pacific Time; `end_date` is inclusive and final analytics data typically lags by two to three days. `data_state: "all"` can include fresher, still-changing rows.
## What can change
| Operation | What happens | Confirmation boundary |
|---|---|---|
| List properties, analytics, sitemaps and URL status | Reads Search Console data | No change |
| Add a property | Adds a property entry; does not verify it | Changes account access |
| Submit or resubmit a sitemap | Requests processing of a sitemap | Changes Search Console state |
| Delete a property | Unlinks the property from the account; Google data is not deleted | Destructive |
| Delete a sitemap | Removes a submitted sitemap | Destructive |
| Raw API request | May call a write or delete endpoint | Potentially destructive |
The AI client controls confirmation prompts. The server marks reads, writes and destructive calls so the client can distinguish an inspection from a real change.
## Getting access
Google Search Console requires OAuth 2.0; an API key is not enough. There are two ways in, and the first one needs no configuration files.
### Connect from the chat (recommended)
Say "connect Google Search Console" and the assistant runs the flow with you:
1. `setup_instructions` prints the checklist: create or select a Google Cloud project, enable **Google Search Console API**, configure the consent screen and create a **Desktop app** OAuth client.
2. Download that client's JSON ("Download JSON") and give the assistant its **path** — `set_client` stores it owner-only. The secret never goes through the conversation.
3. `start_login` returns a Google consent link. Open it **on this machine** and approve; the code comes back to a one-shot listener on `127.0.0.1` (PKCE), never through the chat.
4. `finish_login` exchanges the code and saves the tokens to `~/.config/mcp-google-search-console/credentials.json` (mode 0600) and verifies them with a real Google Search Console API call — so an API that is still switched off is caught right there.
The tokens are re-read on every call, so the connection works immediately — no restart of the AI app. `auth_status` shows what is connected, `logout` revokes and deletes it.
### Environment variables (CI, unattended installs)
1. Create or select a Google Cloud project and enable **Google Search Console API**.
2. Configure the OAuth consent screen and create a **Desktop app** OAuth client.
3. Use the [OAuth 2.0 Playground](https://developers.google.com/oauthplayground) with **Use your own OAuth credentials** to authorize the Google account that can access the properties and obtain a refresh token.
4. Use `https://www.googleapis.com/auth/webmasters` to include sitemaps and property changes. Use `https://www.googleapis.com/auth/webmasters.readonly` only if you intentionally need read-only access.
Testing-mode refresh tokens can expire after seven days. Publish the OAuth app, or use an Internal Workspace app, for long-lived access. Treat the client secret and refresh token as passwords.
## Configuration
Every variable is optional — with none of them the server connects [from the chat](#connect-from-the-chat-recommended).
| Variable | Required | Description |
|---|---|---|
| `GOOGLE_SEARCH_CONSOLE_CLIENT_ID` | No* | OAuth client ID. |
| `GOOGLE_SEARCH_CONSOLE_CLIENT_SECRET` | No* | OAuth client secret. |
| `GOOGLE_SEARCH_CONSOLE_REFRESH_TOKEN` | No* | OAuth refresh token. |
| `GOOGLE_SEARCH_CONSOLE_ACCESS_TOKEN` | No* | Short-lived alternative to the OAuth trio. |
| `GOOGLE_SEARCH_CONSOLE_OAUTH_PORT` | No | Fixed loopback port for the in-chat login; useful over SSH port forwarding. |
| `GOOGLE_SEARCH_CONSOLE_API_BASE` | No | API base URL override. |
| `GOOGLE_SEARCH_CONSOLE_TIMEOUT_MS` | No | Per-request timeout; default `60000` ms. |
| `GOOGLE_SEARCH_CONSOLE_MAX_RETRIES` | No | Temporary-error retries; default `3`. |
\* Provide either the OAuth trio or an access token.
## Data, limits and background work
- **Privacy.** The local server calls Google and sends anonymous telemetry with an installation ID, versions and tool names — never OAuth tokens, property data, tool arguments or prompts. Set `ASKADS_TELEMETRY=0` to opt out.
- **API limits.** URL inspection allows 2,000 inspections per property per day and 600 per minute. Analytics returns at most 25,000 rows per request; long-tail anonymized queries are never returned. Use pagination and do not inspect whole sites URL by URL.
- **No background monitoring.** The server works only while called. If your AI app supports scheduled tasks, it can periodically check a sitemap or an important URL.
## Technical documentation
- [MCP capability catalog](./docs/capabilities/index.md) — task-oriented pages for every tool.
- [All tools and inputs](./docs/TOOLS.md)
- [Development documentation](./docs/DEVELOPMENT.md)
- [Publishing documentation](./docs/PUBLISHING.md)
- [Google Search Console API](https://developers.google.com/webmaster-tools)
## Support
Found a bug or need a scenario? [Create an issue](https://github.com/A1-x-Tech/mcp-google-search-console/issues) or write in [Telegram](https://t.me/a1_mcp).
<br>
<p align="center">
<img src="https://github.com/ztemerbekov/a1-yandex-kit-skills/raw/main/assets/images/mona-hifive-yandex-kit-warm.gif" alt="Две Моны дают пять" width="256">
</p>
<p align="center">
You made it to the end!
</p>
TDQS
Scored across 18 tools
Each tool targets a distinct resource/action: site management, sitemaps, analytics, inspection, auth, and a raw escape hatch. Even search_analytics vs get_top_queries are clearly separated—the latter is an explicit convenience wrapper. No two tools appear to do the same thing.
All tools follow a consistent verb_noun pattern (list_sites, get_site, add_site, delete_site, start_login, submit_sitemap, etc.). Even auth_status and setup_instructions fit the pattern when read as verb+noun. No style mixing or vague verbs.
18 tools is above the typical 3-15 range but justified by the server's broad scope: auth, site management, sitemaps, analytics, URL inspection, and a raw API fallback. Each tool serves a clear purpose with no redundancy; slightly heavy but well-scoped.
The tool surface covers all core Search Console API operations: sites (list/get/add/delete), sitemaps (list/get/submit/delete), search analytics (full query + convenience wrapper), URL inspection, and full OAuth lifecycle. The raw_request tool ensures no endpoint is unreachable, leaving no obvious gaps or dead ends.