creator-seo-mcp
Provides tools to query Google Search Console data for SEO analysis, including striking distance keywords, page performance, content decay detection, cannibalization detection, and revenue-weighted opportunity ranking.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@creator-seo-mcp Find my best SEO opportunities this week."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
creator-seo-mcp
An MCP server for content creators that connects Google Search Console to your page content and ranks every SEO opportunity by estimated revenue, not vanity clicks.
Generic SEO tools surface raw GSC numbers. This one tells you which fix pays most, based on your actual display-ad RPM and affiliate commission rates.
What it does
Six tools, designed to work together in a full creator-SEO workflow:
Tool | Description |
| Queries ranking at positions 4-15 with real impression volume, the ranking page, and the gap to page 1 |
| Full GSC picture for one URL: clicks, impressions, CTR, position, and the queries driving it |
| Flags pages losing clicks or impressions month-over-month |
| Fetches a URL, reads title/meta/headings/body, compares against a target query, and proposes concrete edits |
| Detects queries where multiple pages compete, splitting authority |
| The headline call: combines all signals, weights by revenue, returns a single ranked action list |
Related MCP server: Google Search Console MCP Server
Install
uvx creator-seo-mcpOr with pip:
pip install creator-seo-mcpSetup
1. Google Search Console credentials
Follow docs/gsc-setup.md to:
Enable the Search Console API in Google Cloud
Create an OAuth 2.0 Desktop client
Download
credentials.json
2. Claude Desktop config
Add to your claude_desktop_config.json:
{
"mcpServers": {
"creator-seo-mcp": {
"command": "uvx",
"args": ["creator-seo-mcp"],
"env": {
"GOOGLE_CREDENTIALS_PATH": "/path/to/credentials.json",
"CREATOR_SEO_SITE_RPM": "15"
}
}
}
}3. Claude Code config
claude mcp add creator-seo-mcp uvx creator-seo-mcp \
-e GOOGLE_CREDENTIALS_PATH=/path/to/credentials.json \
-e CREATOR_SEO_SITE_RPM=154. Revenue config (optional but recommended)
Set your display-ad RPM so opportunities are ranked by real dollars:
export CREATOR_SEO_SITE_RPM=22 # your Mediavine/Raptive RPM
export CREATOR_SEO_AFFILIATE_CATEGORIES='{"gaming-deals": 2.0}'See .env.example for all options.
Example agent prompts
"Show me my top five revenue-weighted SEO opportunities for this week."
"Which of my posts are losing traffic compared to last month?"
"My Baldur's Gate 3 build guide is stuck on page 2. Audit it against its main keyword and tell me what to fix."
"Find any posts that are competing with each other for the same search term."
"What is the on-page gap between this article and the query it is trying to rank for?"
Two Average Gamers case study
Real output from get_top_opportunities, run live against Two Average Gamers (a top-200 gaming blog on Mediavine display ads and Amazon affiliate links) on 2026-08-25, over the trailing 28 days (392K impressions, 3.5K clicks per Search Console).
This is the actual "before" baseline, not a mockup. Dollar estimates use the default $15 RPM placeholder (set CREATOR_SEO_SITE_RPM to your real network RPM for accurate figures). Out of 147 scored opportunities across all three signal types, the top 5 by estimated revenue:
# | Type | Page | Query | Finding |
1 | Striking distance | "letterboxd for games" | Position 4.6, 97K impressions, 1.6-position gap to page 1 | |
2 | Cannibalization | same page + 3 others | "letterboxd for games" | 4 TAG pages splitting 99.8K impressions on the same query |
3 | Striking distance | "backloggd" | Position 5.5, 11.4K impressions, 2.5-position gap to page 1 | |
4 | Cannibalization | same page + 2 others | "backloggd" | 3 TAG pages splitting 15.1K impressions on the same query |
5 | Striking distance | "backloggd" | Position 5.5, 3.7K impressions, 2.5-position gap to page 1 |
analyze_content_decay also flagged 40 pages losing clicks month-over-month on this run, the worst being a 100% drop (10 clicks to 0) on a Palworld lawsuit post and a 61% drop on a Palworld tower-boss guide, both worth a content refresh before the traffic they had is gone for good.
Update (2026-08-25): the title/H1 edits for the two striking-distance opportunities above are live. Re-running audit_page_onpage against the real pages confirms it:
Page | Target query | In title? | In H1? | Remaining suggestions |
"letterboxd for games" | ✅ | ✅ | none | |
"backloggd" | ✅ | ✅ | none | |
"backloggd" | ✅ | ✅ | none |
No position or traffic movement to report yet, and that's expected, not a null result: Search Console's data has a 2-3 day freshness lag, so a same-day re-pull still covers the identical 2026-07-26 to 2026-08-22 window as the original baseline above, before any of these edits existed. Google also needs to re-crawl and re-rank the pages, which typically takes longer than the raw data lag. This section will be updated again once a GSC pull actually covers the post-edit period (realistically 2-3 weeks out) with the real position and click change, not before.
Contributing
Issues and PRs welcome. See CONTRIBUTING.md for guidelines.
License
MIT
Available Tools
8 toolsanalyze_content_decayA
Compare recent vs prior period for every page and flag declining traffic. Results are sorted worst decliners first (most negative percent change at top).
Inputs: site_url (required), min_clicks, recent_days, compare_days. Returns: result
audit_page_onpageB
Fetch a page, parse its title/meta/headings/body, and return concrete edit suggestions for the target query.
Inputs: page_url (required), target_query (required). Returns: h1, url, title, headings, +6 more
find_cannibalizationB
Find queries where multiple pages are competing, splitting authority and confusing Google. Results are sorted by total impression volume.
Inputs: site_url (required), days, min_impressions. Returns: result
get_page_performanceA
Full GSC picture for one URL: clicks, impressions, CTR, position, and top queries.
Inputs: page_url (required), site_url (required), days. Returns: ctr, page, clicks, position, +2 more
get_striking_distance_keywordsA
Find queries ranking in striking distance of page 1 (positions 4-15 by default). Results are sorted by impression volume so the highest-potential opportunities come first.
Inputs: site_url (required), days, limit, position_max, +2 more. Returns: result
get_top_opportunitiesA
The headline call. Combines striking-distance keywords, content decay, and cannibalization, weights each by estimated revenue, and returns a single ranked action list. Revenue estimates use your configured RPM and affiliate multipliers.
Inputs: site_url (required), days, limit, site_rpm. Returns: result
health_checkA
Check that the server is running and list your verified GSC properties.
Inputs: no named parameters. Returns: status, version, token_status, verified_sites
list_sitesA
List all Google Search Console properties you have access to.
Inputs: no named parameters. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 8 tools
Most tools have distinct purposes (e.g., keyword distance, performance, decay, audit, cannibalization). Only health_check and list_sites overlap slightly as both list GSC properties, but health_check primarily serves as a status check.
Tools generally follow a verb_noun pattern (list_sites, get_striking_distance_keywords, analyze_content_decay). health_check deviates (noun_noun) and 'audit_page_onpage' uses an awkward preposition, but overall pattern is mostly consistent.
8 tools cover the core SEO workflow (health, sites, opportunities, performance, decay, audit, cannibalization, and a combined report) without being excessive. The scope is well-calibrated for creator SEO use.
The set covers key analysis areas: opportunity identification, page performance, content decay, on-page audit, and cannibalization. Missing are execution tools (e.g., update page) and broader keyword research, but the core analytical workflow is complete.
Maintenance
Related MCP Connectors
Turn Search Console data into SEO actions, content, publishing, indexing, and AI insights.
Turns SEO insight into page changes: keyword research, SERP and rank data, rewrites you approve.
AI SEO agent on your Google Search Console data: audits, quick wins, content, WordPress publishing.
SEO data your AI can cite: Search Console, GA4, keywords, backlinks, SERPs and AI visibility.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceEnables querying Google Search Console and Google Analytics 4 through natural language, with tools for SEO analysis like anomaly detection, cannibalization detection, and opportunity scoring.1MIT
- AlicenseBqualityDmaintenanceConnects Google Search Console to AI assistants, enabling SEO data analysis through natural language conversations.32MIT
- AlicenseNot gradedqualityDmaintenanceConnects Google Search Console to AI assistants, enabling SEO data analysis via natural language conversations.MIT
- AlicenseBqualityBmaintenanceEnables users to query Google Search Console data and rank striking-distance keywords, decaying pages, and cannibalized queries by estimated revenue rather than clicks. It provides read-only, local, headless-safe access with auditable dollar estimates and optional real RPM data.9MIT