Onpage SEO involves constantly switching between different tools to perform specific tasks. This manual workflow got me thinking: a significant portion of these tasks could be safely automated by combining AI with SEO tool APIs.
After testing several approaches, I came up with a solution that consistently produces good results: pre-defined workflows following highly specific prompts.
I’ll walk you through two workflows I’ve been using over the past few months. The first automates keyword research based on an article title and predefined content structure. The second helps optimize underperforming content.
At the end, I’ll show you how to clone my GitHub repo and set up these workflows yourself.
What you’ll need:
- Claude Desktop
- MCP
- KeywordsEverywhere API key
- SERP API key
Let’s get started.
Topics
TL;DR
I demonstrate two customized MCP workflows that significantly speed up on-page SEO by automating repetitive tasks using AI and SEO tool APIs. The first workflow automates keyword research based on article structure, while the second optimizes underperforming content.
By following the provided steps, you’ll learn how to set up these workflows using Claude Desktop, MCP, KeywordsEverywhere API, and SERP API. These workflows not only save time but also provide actionable recommendations for optimizing article structure, content, and technical aspects.
With these workflows, you’ll be able to create well-optimized content more efficiently, ultimately improving your SEO results. The best part is that you can customize these workflows to fit your specific needs and use other SEO APIs if preferred.
Workflow 1: Onpage keyword research
Purpose: Structure-based keyword research and SERP optimization for new content creation.
Best for: Blog posts, content marketing, planning new articles
How does the workflow work
Step 1: Article Structure Analysis
Takes your article idea (title, heading structure, content outline) and extracts:
- Core topic identification and main theme
- Content intent analysis (educational, commercial, informational)
- Target audience level assessment
- Content type classification (how-to, listicle, comparison, guide)
Step 2: Semantic Topic Discovery & Keyword Research
Uses Keywords Everywhere API to:
- Extract 3-5 seed keywords from your topic analysis
- Run
related_keywords_discovery
for semantic keyword expansion - Execute
keyword_research_analysis
for volume/competition data - Apply
keyword_opportunity_scorer
to prioritize targets - Map semantic relationships between topic clusters
Step 3: SERP Analysis & Intent Classification
Uses SerpAPI to:
- Run
classify_search_intent_data
to understand what Google expects - Execute
serp_data_collector
for the current SERP landscape analysis - Analyze SERP features (featured snippets, PAA, related searches)
- Map content format preferences from top results
Step 4: Content Structure Optimization
Generates specific recommendations:
- Optimized article structure with strategic heading placement
- Keyword integration strategy across H1-H6 hierarchy
- Content sections to add based on competitor gap analysis
- SERP feature targeting (FAQ sections for PAA opportunities)
Step 5: Implementation Checklist
Provides an actionable roadmap:
- Title optimization with primary keyword placement
- Heading restructure based on SERP insights
- Content length targets vs competitors
- Internal linking strategy
- Schema markup recommendations
Example: “How to Start a Home Garden” Article
Input Structure:
H1: "How to Start a Home Garden"
H2: "Choosing Your Garden Location"
H2: "Essential Gardening Tools"
H2: "Best Plants for Beginners"
H2: "Watering and Care Tips"
Workflow Output:
SEMANTIC ANALYSIS RESULTS:
Core Topic Clusters Identified:
- Garden planning and setup (primary)
- Plant selection and tool requirements (supporting)
- Maintenance routines and care (supporting)
- Seasonal timing and space planning (missing)
KEYWORD RESEARCH FINDINGS:
Primary: "how to start a garden" (12,100 volume, medium competition)
Semantic Variations: "beginner garden setup" (890 volume), "garden planning guide" (720 volume)
Opportunity Keywords: "container gardening for beginners" (1,200 volume, low competition)
Long-tail Targets: "when to start a vegetable garden" (480 volume, very low competition)
SERP ANALYSIS INSIGHTS:
- 70% of top results include seasonal timing sections (missing from current structure)
- Budget considerations appear in 60% of top content (gap opportunity)
- Garden types (container, raised bed, in-ground) covered by all top 5 results
- FAQ sections present in 80% of ranking content
OPTIMIZED STRUCTURE RECOMMENDATIONS:
H1: How to Start a Garden: Complete Beginner's Guide
H2: Planning Your Garden Layout and Location
H2: Garden Types: Container vs Raised Bed vs In-Ground
H2: When to Start Your Garden (Seasonal Timing Guide)
H2: Essential Gardening Tools for Beginners
H2: Best Plants for New Gardeners
H2: Budget-Friendly Garden Setup Tips
H2: Watering, Care, and Maintenance Schedule
H2: Common Beginner Mistakes to Avoid (FAQ)
IMPLEMENTATION CHECKLIST:
□ Add seasonal timing section (competitor gap)
□ Include garden types comparison (SERP requirement)
□ Create FAQ section targeting PAA questions
□ Integrate "container gardening" keyword cluster
□ Add budget considerations throughout
□ Target 2,500-3,000 words (competitor average: 2,800)
Workflow 2: Existing Content Optimization
Purpose: Improves underperforming content into top-ranking pages through comprehensive analysis.
Best for: Optimizing existing blog posts, landing pages, product pages that aren’t ranking
How does the workflow work
Phase 1: Deep Current State Analysis
scrape_seo_data
→ Extract all content and technical elementspage_speed_metrics
(desktop + mobile) → Performance baselinevalidate_structured_data
→ Schema analysis and rich results eligibilitycollect_server_headers
→ Technical infrastructure auditdetect_javascript_rendering
→ Crawlability assessment
Phase 2: Competitive Landscape & Intent Analysis
classify_search_intent_data
→ Intent classification for primary keywordserp_data_collector
→ Current SERP landscape mappinganalyze_serp_content_alignment
→ Content gap analysis vs top 10 resultsanalyze_serp_feature_opportunities
→ SERP feature targeting strategy
Phase 3: Semantic & Keyword Universe Expansion
related_keywords_discovery
→ Semantic keyword expansionkeyword_research_analysis
→ Volume/competition data for expanded listkeyword_opportunity_scorer
→ Strategic prioritization- Multiple rounds of semantic mapping for related entities and concepts
Phase 4: Advanced Content Gap Analysis
scrape_seo_data
on top 3-5 competitorspage_speed_metrics
for performance benchmarking- Content depth assessment and topic coverage analysis
- Authority signal identification and competitive advantage mapping
Phase 5: NLP & Content Quality Analysis
- Semantic analysis (entity coverage, topic modeling, keyword density)
- Readability assessment (Flesch scores, sentence structure, paragraph length)
- Authority signals audit (expertise, citations, freshness indicators)
- Conversion optimization analysis (user intent alignment, CTA placement)
Phase 6: Comprehensive Optimization Strategy
- Technical enhancement recommendations (Core Web Vitals, schema, mobile)
- Content transformation strategy (semantic enrichment, structure optimization)
- Authority building plan (expert sources, original research, social proof)
- Conversion optimization (intent alignment, user experience improvements)
Example: Underperforming “Email Marketing Automation” Page
Current State Analysis:
COMPETITIVE ANALYSIS:
Top 10 Competitor Patterns:
- 90% include automation workflow examples (missing from our content
- 80% have tool comparison sections (we have brief mentions only)
- 70% include ROI/metrics sections (completely missing)
- 60% have video content embedded (we have none)
- Average content depth: 3,800 words with 15+ subsections
Keyword Universe Expansion:
SEMANTIC KEYWORD ANALYSIS:
Primary Cluster: "email marketing automation" (8,100 volume)
Supporting Keywords Discovered:
- "marketing automation workflows" (1,200 volume, medium comp)
- "email sequence automation" (890 volume, low comp)
- "automated email campaigns" (2,400 volume, high comp)
- "drip campaign automation" (720 volume, medium comp)
Gap Analysis vs Current Content:
Missing 23 related keywords that competitors rank for:
- "email automation tools comparison" (650 volume)
- "automated email marketing ROI" (380 volume)
- "email workflow templates" (540 volume)
Comprehensive Optimization Strategy:
CRITICAL PRIORITY FIXES (Week 1-2):
□ Add FAQ schema targeting 8 PAA questions
□ Implement lazy loading for 1.5s LCP improvement
□ Restructure H2/H3 hierarchy with semantic keywords
□ Add internal links to 5 related automation articles
HIGH PRIORITY CONTENT (Week 3-6):
□ Add 1,200-word "Automation Workflow Design" section
□ Create tool comparison table (HubSpot vs Mailchimp vs ActiveCampaign)
□ Include ROI calculator and metrics tracking section
□ Add 8 workflow template examples with screenshots
STRATEGIC DIFFERENTIATION (Week 7-12):
□ Embed video walkthrough of automation setup
□ Add interactive workflow builder tool
□ Include original survey data on automation ROI
□ Create downloadable workflow template library
KEYWORD INTEGRATION PLAN:
- "marketing automation workflows" → New H2 section
- "email sequence automation" → Dedicated subsection
- "automated email campaigns" → Case study examples
- "drip campaign automation" → FAQ section targeting
Both workflows connect to SEO APIs for data collection, then Claude applies optimization logic following the workflow specified in the prompts to generate recommendations.
How to set it up
Step 1: Clone the Repository
git clone https://github.com/dexter480/mcp-seo-workflows
cd seo-workflows-claude
pip install -r requirements.txt
Step 2: Get Free API Keys
- Keywords Everywhere: keywordseverywhere.com – 100 searches/month free
- SerpAPI: serpapi.com – 100 searches/month free
Step 3: Configure Environment Variables
cp .env.example .env
Edit the .env
file with your API keys:
KEYWORDS_EVERYWHERE_API_KEY=your_key_here
SERPAPI_KEY=your_key_here
GOOGLE_PAGESPEED_KEY=your_key_here
Step 4: Connect to Claude Desktop
Find your Claude Desktop config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json
- Windows:
%APPDATA%/Claude/claude_desktop_config.json
Add this configuration:
{
"mcpServers": {
"seo-analyzer": {
"command": "python",
"args": ["/full/path/to/your/seo_scraper_mcp.py"]
}
}
}
Important: Replace /full/path/to/your/
with the actual absolute path to where you cloned the repository.
Step 5: Test the Setup
- Restart Claude Desktop
- Try this test prompt:
Use the scrape_seo_data tool to analyze https://google.com
If you get SEO data back, you’re ready to use the workflows!
Flexibility
The best part of these workflows is the flexibility. Once you understand how it works, you’ll realize that you can automate a ton of on-page tasks, some more reliably than others.
You also don’t have to use the APIs I used here. You can use DataForSEO or Ahrefs API, there are no rules.