seo-mcp
MCP server that gives AI assistants live access to Google Search Console and Bing Webmaster Tools for search performance, indexing, keyword research, and crawl health analysis directly in the chat.
README
@patchwindow/seo-mcp
MCP server for Google Search Console and Bing Webmaster Tools. Give your AI assistant live access to search performance data, indexing status, keyword research, crawl health, and more — directly in the chat.
Features
- 10 SEO tools across Google Search Console and Bing Webmaster Tools
- Keyword research via Bing — search volume and related keywords not available in GSC
- Traffic drop analysis — automatically compare two periods and surface the biggest drops
- Striking distance finder — identify queries in positions 4–20 ready for quick ranking gains
- Brand vs. non-brand split — segment traffic without leaving your AI chat
- Dual URL inspection — inspect any URL in both Google and Bing simultaneously
- Simple auth — Bing needs one env var; GSC uses a one-time OAuth2 flow with auto-refreshing tokens
- Config file defaults — set your site URL once, skip it on every tool call
Installation
Install globally:
npm install -g @patchwindow/seo-mcp
Or run without installing:
npx @patchwindow/seo-mcp
Setup
Bing Webmaster Tools
Bing uses an API key. No OAuth required.
- Go to Bing Webmaster Tools
- Navigate to Settings → API Access
- Generate an API key
- Set the environment variable:
export BING_WEBMASTER_API_KEY="your-api-key"
That's it.
Google Search Console
GSC requires OAuth2. Service accounts do not work with the Search Console API — user credentials are required.
Step 1: Create a Google Cloud project
- Go to Google Cloud Console
- Create a new project (or select an existing one)
- Go to APIs & Services → Library
- Search for Google Search Console API and enable it
Step 2: Create OAuth2 credentials
- Go to APIs & Services → Credentials → Create Credentials → OAuth 2.0 Client ID
- Choose Web application as the application type
- Under Authorized redirect URIs, add:
http://localhost:3847/callback - Copy the Client ID and Client Secret
Step 3: Authenticate (one-time)
Set your credentials as environment variables, then run the auth command:
export GSC_CLIENT_ID="your-client-id.apps.googleusercontent.com"
export GSC_CLIENT_SECRET="your-client-secret"
npx @patchwindow/seo-mcp auth gsc
A browser window opens for Google login. After approving, the token is saved to ~/.seo-mcp/gsc-token.json. You only need to do this once — the token refreshes automatically.
Config File (optional)
Create ~/.seo-mcp/config.json to set default site URLs. This lets you skip the site_url parameter on every tool call:
{
"gsc": {
"default_site": "sc-domain:example.com"
},
"bing": {
"default_site": "https://example.com/"
},
"output": {
"max_rows": 500
}
}
GSC site URL format: Use sc-domain:example.com for domain properties or https://example.com/ for URL prefix properties. Check which type you have in Search Console under your property settings.
Claude Desktop Integration
Add the following to your claude_desktop_config.json (typically at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS or %APPDATA%\Claude\claude_desktop_config.json on Windows):
{
"mcpServers": {
"seo": {
"command": "npx",
"args": ["@patchwindow/seo-mcp"],
"env": {
"BING_WEBMASTER_API_KEY": "your-bing-api-key",
"GSC_CLIENT_ID": "your-client-id.apps.googleusercontent.com",
"GSC_CLIENT_SECRET": "your-client-secret"
}
}
}
}
Cursor / Windsurf (.cursor/mcp.json or mcp.json):
{
"seo-mcp": {
"command": "npx",
"args": ["@patchwindow/seo-mcp"],
"env": {
"BING_WEBMASTER_API_KEY": "your-bing-api-key",
"GSC_CLIENT_ID": "your-client-id.apps.googleusercontent.com",
"GSC_CLIENT_SECRET": "your-client-secret"
}
}
}
Restart your AI client after saving the config.
Tools Reference
* = not required if default_site is set in ~/.seo-mcp/config.json.
Google Search Console
| Tool | Description | Required params | Optional params |
|---|---|---|---|
gsc_search_performance |
Clicks, impressions, CTR, and position. Supports grouping by query, page, country, device, or date. | start_date, end_date |
site_url*, dimensions, filter_query, filter_page, filter_device, filter_country, row_limit |
gsc_striking_distance |
Queries in positions 4–20 sorted by impressions — best candidates for quick ranking improvements. | start_date, end_date |
site_url*, min_position, max_position, min_impressions, row_limit |
gsc_traffic_drop |
Compare two date periods and surface pages or queries with the largest click drops. | current_start, current_end, previous_start, previous_end |
site_url*, dimension, min_drop_percent, min_clicks_previous, row_limit |
gsc_url_inspection |
Indexing status, crawl date, canonical URL, page fetch state, rich results, and mobile usability for a specific URL. | url |
site_url* |
gsc_sitemap_list |
All sitemaps submitted to GSC with status, URL counts, indexed counts, and error summary. | — | site_url* |
gsc_brand_nonbrand |
Split search traffic into branded and non-branded segments with aggregated clicks, impressions, CTR, and position. | start_date, end_date, brand_terms |
site_url*, show_top_queries |
Bing Webmaster Tools
| Tool | Description | Required params | Optional params |
|---|---|---|---|
bing_keyword_research |
Monthly search volume and related keywords from Bing. This data is exclusive to Bing — not available in GSC. | keyword |
site_url*, country, language, include_related |
bing_crawl_health |
Crawl statistics (total crawled, errors by type: 4xx, timeouts, DNS failures, blocked) and a list of specific crawl issues. | — | site_url*, show_issues, max_issues |
bing_url_inspection |
HTTP status, indexing state, crawl date, page title, internal/external link counts, and redirect target for a URL. | url |
site_url* |
bing_sitemap_list |
Sitemaps registered in Bing Webmaster Tools with URL counts, indexed counts, errors, and last crawl time. | — | site_url* |
Parameter details
<details> <summary>gsc_search_performance</summary>
| Parameter | Type | Description |
|---|---|---|
site_url |
string | GSC site URL (sc-domain:example.com or https://example.com/) |
start_date |
string | Start date, YYYY-MM-DD |
end_date |
string | End date, YYYY-MM-DD |
dimensions |
array | Group by: query, page, country, device, date. Default: ["query"] |
filter_query |
string | Filter to queries containing this string |
filter_page |
string | Filter to this exact page URL |
filter_device |
string | DESKTOP, MOBILE, or TABLET |
filter_country |
string | ISO 3166-1 alpha-3 country code, e.g. USA |
row_limit |
number | Max rows. Default: 100. Max: 25000 |
</details>
<details> <summary>gsc_striking_distance</summary>
| Parameter | Type | Description |
|---|---|---|
site_url |
string | GSC site URL |
start_date |
string | Start date, YYYY-MM-DD |
end_date |
string | End date, YYYY-MM-DD |
min_position |
number | Min position to include. Default: 4 |
max_position |
number | Max position to include. Default: 20 |
min_impressions |
number | Min impressions to include. Default: 10 |
row_limit |
number | Max results. Default: 50 |
</details>
<details> <summary>gsc_traffic_drop</summary>
| Parameter | Type | Description |
|---|---|---|
site_url |
string | GSC site URL |
current_start |
string | Current period start, YYYY-MM-DD |
current_end |
string | Current period end, YYYY-MM-DD |
previous_start |
string | Previous period start, YYYY-MM-DD |
previous_end |
string | Previous period end, YYYY-MM-DD |
dimension |
string | page or query. Default: page |
min_drop_percent |
number | Minimum drop % to flag. Default: 20 |
min_clicks_previous |
number | Minimum clicks in previous period. Default: 5 |
row_limit |
number | Max results. Default: 25 |
</details>
<details> <summary>gsc_brand_nonbrand</summary>
| Parameter | Type | Description |
|---|---|---|
site_url |
string | GSC site URL |
start_date |
string | Start date, YYYY-MM-DD |
end_date |
string | End date, YYYY-MM-DD |
brand_terms |
array | Brand terms to match, case-insensitive. E.g. ["acme", "acmecorp"] |
show_top_queries |
boolean | Include top 10 queries per segment. Default: true |
</details>
<details> <summary>bing_keyword_research</summary>
| Parameter | Type | Description |
|---|---|---|
keyword |
string | The keyword to research |
site_url |
string | Bing site URL |
country |
string | Two-letter country code, e.g. US. Default: US |
language |
string | Language code, e.g. en-US. Default: en-US |
include_related |
boolean | Include related keywords. Default: true |
</details>
<details> <summary>bing_crawl_health</summary>
| Parameter | Type | Description |
|---|---|---|
site_url |
string | Bing site URL |
show_issues |
boolean | Include specific crawl issues. Default: true |
max_issues |
number | Max issues to show. Default: 20 |
</details>
Examples
Ask your AI assistant:
Find quick wins for content optimization:
"Show me my top striking distance queries for the last 90 days. I want queries where I'm ranking between 5 and 15 with at least 50 impressions."
Diagnose a traffic drop:
"Compare my search traffic for April 2025 vs April 2024 by page. Flag anything that dropped more than 30%."
Research a topic before writing:
"Research the keyword 'typescript performance' in Bing. Show me monthly search volume and the top related keywords."
Audit indexing:
"Inspect the URL https://example.com/blog/my-post and tell me its indexing status, canonical URL, and whether it's eligible for rich results."
Understand brand exposure:
"Split my search traffic for Q1 2025 into branded and non-branded. My brand terms are 'acme' and 'acmecorp'. Show me the top 10 queries for each segment."
Roadmap
Planned for v2:
gsc_top_pages— top pages by traffic with trend comparisonbing_link_counts— inbound link data (not available in GSC API)bing_url_submit— trigger reindexing for new or updated URLsgsc_batch_url_inspection— inspect a list of URLs with rate-limit handlingbing_crawl_settings— read and update Bing crawl frequency settings- Multi-site support — switch between properties without reconfiguring
Building from Source
git clone https://github.com/patchwindow/seo-mcp
cd seo-mcp
npm install
npm run build
Contributing
Issues and pull requests are welcome. Please open an issue before starting work on a significant change.
License
MIT — Patch Window
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.