Skip to content

Free, open source blog writer for Claude Code. Researches current sources, writes answer-first posts for Google and AI answers, has an independent reviewer score each one out of 100 and ships only after five delivery gates. 32 skills, 30 /blog commands, 5 agents. Active development at AI-Marketing-Hub/claude-blog; public releases ship here.

20.2 KB
Raw Blame

MCP Integration Guide

Optional Model Context Protocol (MCP) server integrations that extend claude-blog with live data from SEO platforms, analytics services, and performance monitoring tools.

Important: claude-blog works fully without any MCP servers. These integrations are optional enhancements for teams that already use these platforms.


Overview

                    +---------------------------+
                    |      claude-blog          |
                    |    /blog commands          |
                    +------+----+----+----------+
                           |    |    |
              +------------+    |    +------------+
              |                 |                  |
              v                 v                  v
  +-----------------+  +----------------+  +------------------+
  | DataForSEO MCP  |  | Individual     |  | Custom MCP       |
  | (Recommended)   |  | MCP Servers    |  | Servers          |
  |                 |  |                |  |                  |
  | - SERP data     |  | - GSC          |  | - Analytics      |
  | - Keywords      |  | - Ahrefs       |  | - CMS APIs       |
  | - Backlinks     |  | - Semrush      |  | - Custom data    |
  | - On-page       |  | - PageSpeed    |  |                  |
  | - Domain data   |  |                |  |                  |
  | - Content       |  |                |  |                  |
  | - AI Optim.     |  |                |  |                  |
  +-----------------+  +----------------+  +------------------+

Nano Banana MCP - AI Image Generation

The nanobanana-mcp server enables AI image generation within blog workflows. When configured, /blog write and /blog rewrite can generate custom hero images, inline illustrations, and social preview cards via Gemini, in addition to stock photo sourcing from Pixabay/Unsplash/Pexels.

What It Enables

Feature Without nanobanana-mcp With nanobanana-mcp
Hero/cover images Stock photos only Stock photos + AI-generated custom images
Inline illustrations Stock photos only Stock photos + topic-specific AI illustrations
OG/social cards Stock photo crop Custom AI-generated social preview images
Image editing Not available Edit existing blog images (crop, enhance, restyle)
Standalone use Not available /blog image generate <idea> for any image need

Configuration

The setup script writes to Claude Code's user-private global settings by default (safer; never reaches git). To use project-local .mcp.json, opt in with --project (the script will refuse to write a literal key into a tracked file).

# Recommended: writes ~/.claude/settings.json (user-private, mode 0600)
python3 skills/blog-image/scripts/setup_image_mcp.py --key-file /path/to/key.txt   # or omit for an interactive prompt

# Project-local (env-expansion only, NEVER stores literal key)
python3 skills/blog-image/scripts/setup_image_mcp.py --key-file /path/to/key.txt --project
# Then export the key in your shell so the env-expansion resolves:
export GOOGLE_AI_API_KEY="your-key-from-aistudio.google.com"

Get a free API key at: https://aistudio.google.com/apikey

Security: .mcp.json is gitignored as of 2026-04-27. A tracked .mcp.example.json template lives alongside. The setup script pins the nanobanana-mcp package version (@ycse/[email protected]) to mitigate supply-chain risk on npx -y auto-update.

Verify Setup

python3 skills/blog-image/scripts/validate_image_setup.py

Requirements

  • Node.js 18+ (for npx)
  • Google AI API key with access to the selected model. Quotas vary by model, project and tier; check active limits in AI Studio before a run (reviewed 2026-10-07).

DataForSEO MCP (Recommended)

DataForSEO is the recommended MCP integration for claude-blog. It provides a single unified API covering SERP data, keyword research, backlink analysis, on-page auditing, domain analytics, content analysis, and AI optimization, consolidating several third-party research workflows. Search Console owner data, GA4 analytics, and PageSpeed measurements remain distinct inputs with their own access and semantics.

What It Enables

Feature Without DataForSEO MCP With DataForSEO MCP
SERP analysis WebSearch only Live Google/Bing/Yahoo SERP with all features (AI Overviews, PAA, etc.)
Keyword research Not available Search volume, CPC, competition, keyword difficulty, search intent
Backlink analysis Not available Referring domains, anchor text, spam score, new/lost links
On-page auditing Manual review Automated crawl, meta tags, Core Web Vitals, Lighthouse scores
Domain analytics Not available Technology detection, WHOIS data, competitor domain analysis
Content analysis Quality scoring only Sentiment analysis, keyword density, content quality scoring
AI optimization GEO content audit LLM mention tracking, ChatGPT scraping, AI visibility metrics
Competitor research WebSearch only Ranked keywords, traffic estimation, content gap analysis

Enhanced Workflows

/blog brief with DataForSEO data:

Content briefs include real keyword metrics from DataForSEO Labs:

Target Keywords (DataForSEO Labs)
- Primary: "kubernetes monitoring" - 2,400/mo, KD 45, Intent: informational
- Secondary: "k8s observability" - 890/mo, KD 32, Intent: informational
- Question: "how to monitor kubernetes" - 720/mo, KD 28

Competitor SERP Analysis
| Position | Domain              | Backlinks | Word Count |
|----------|---------------------|-----------|------------|
| #1       | competitor-a.com    | 142       | 3,200      |
| #2       | competitor-b.com    | 89        | 2,800      |
| #3       | competitor-c.com    | 67        | 4,100      |

/blog strategy with DataForSEO data:

Strategy documents gain competitive intelligence:

Domain Comparison (DataForSEO Labs)
| Domain           | Organic Traffic | Keywords | Backlinks | Rank |
|------------------|----------------|----------|-----------|------|
| competitor-a.com | 45,000/mo      | 2,340    | 12,400    | 52   |
| competitor-b.com | 28,000/mo      | 1,200    | 8,900     | 47   |
| your-site.com    | 3,200/mo       | 380      | 1,100     | 31   |

Content Gap: 234 keywords where competitors rank but you don't

/blog geo with DataForSEO AI Optimization data:

AI citation audits include LLM visibility metrics:

AI Visibility Report (DataForSEO AI Optimization)
| Metric              | Value | Notes                    |
|---------------------|-------|--------------------------|
| LLM mentions        | 12    | Across ChatGPT, Gemini  |
| AI Overview present | Yes   | For 3/5 target keywords  |
| Brand sentiment     | 0.72  | Positive                 |
| Citation URLs       | 4     | Directly cited pages     |

Configuration

One-command install (recommended):

Load credentials from a local secret file first so they are not written inline:

set -a
. ~/.config/dataforseo.env
set +a
claude mcp add dataforseo \
  --env DATAFORSEO_USERNAME="${DATAFORSEO_USERNAME}" \
  --env DATAFORSEO_PASSWORD="${DATAFORSEO_PASSWORD}" \
  -- npx -y dataforseo-mcp-server

Or remote server (no local install needed):

DATAFORSEO_BASIC_AUTH="$(printf '%s:%s' "$DATAFORSEO_USERNAME" "$DATAFORSEO_PASSWORD" | base64)"
claude mcp add --transport http dataforseo https://mcp.dataforseo.com/http \
  --header "Authorization: Basic ${DATAFORSEO_BASIC_AUTH}"

Or add to ~/.claude/settings.json manually:

{
  "mcpServers": {
    "dataforseo": {
      "command": "npx",
      "args": ["-y", "dataforseo-mcp-server"],
      "env": {
        "DATAFORSEO_USERNAME": "${DATAFORSEO_USERNAME}",
        "DATAFORSEO_PASSWORD": "${DATAFORSEO_PASSWORD}",
        "ENABLED_MODULES": "SERP,KEYWORDS_DATA,ONPAGE,DATAFORSEO_LABS,BACKLINKS,DOMAIN_ANALYTICS,BUSINESS_DATA,CONTENT_ANALYSIS,AI_OPTIMIZATION"
      }
    }
  }
}

Best Practices

  1. Store credentials in environment variables (not in settings.json):

    # Store DATAFORSEO_USERNAME and DATAFORSEO_PASSWORD in this 0600 file.
    chmod 600 ~/.config/dataforseo.env
    set -a
    . ~/.config/dataforseo.env
    set +a
    
  2. Use field filtering to reduce token usage (~75%): Create a field config JSON file and set FIELD_CONFIG_PATH:

    {
      "env": {
        "FIELD_CONFIG_PATH": "/path/to/dataforseo-field-config.json"
      }
    }
    

    A comprehensive field config is available in the repository at skills/seo/dataforseo-field-config.json (if using the companion /seo skill).

  3. Enable only the modules you need via ENABLED_MODULES to reduce available tools and improve response relevance.

Available Modules

Module What It Provides
SERP Live Google/Bing/Yahoo search results with all SERP features
KEYWORDS_DATA Search volume, CPC, competition from Google Ads
DATAFORSEO_LABS Keyword research, domain analysis, competitor data
BACKLINKS Backlink profiles, referring domains, anchor text
ONPAGE Website crawling, meta analysis, Core Web Vitals, Lighthouse
DOMAIN_ANALYTICS Technology detection, WHOIS records
BUSINESS_DATA Google Maps listings, reviews, business info
CONTENT_ANALYSIS Brand citations, sentiment analysis, phrase trends
AI_OPTIMIZATION LLM mention tracking, ChatGPT scraping, AI keyword discovery

Setup Requirements

  1. DataForSEO account with API credentials
  2. Node.js 18+ (for npx dataforseo-mcp-server)
  3. API username and password from the DataForSEO dashboard

Alternative Individual MCP Integrations

The following individual MCP servers can be used instead of (or alongside) DataForSEO for teams that already have accounts with these platforms.

Google Search Console MCP

What It Enables

Feature Without GSC MCP With GSC MCP
Content decay detection Manual check Automated: flags posts with 20%+ QoQ traffic decline
Keyword tracking Not available Live keyword rankings and CTR data
Query analysis Not available Actual search queries driving traffic
AI Overview impact Not available CTR changes when AI Overviews appear
Freshness scheduling Time-based (30 days) Data-driven (based on actual performance drops)

Enhanced Workflows

/blog audit with GSC data:

Instead of scoring posts only on content quality, the audit can incorporate actual performance data:

Blog Audit: 47 Posts Analyzed

| Post              | Quality | Traffic (QoQ) | CTR    | Action          |
|-------------------|---------|---------------|--------|-----------------|
| ai-search-guide   | 85/100  | -35%          | 1.2%   | Content decay!  |
| kubernetes-setup   | 72/100  | +12%          | 3.1%   | Quality fixes   |
| react-patterns     | 91/100  | -5%           | 4.2%   | Monitor         |

/blog calendar with GSC data:

Editorial calendars can prioritize updates based on actual traffic decay rather than arbitrary 30-day cycles:

Freshness Update Queue (Data-Driven)
| Post              | Last Updated | Traffic Change | Priority |
|-------------------|-------------|----------------|----------|
| ai-search-guide   | 45 days ago | -35% QoQ       | Critical |
| seo-strategy      | 60 days ago | -22% QoQ       | High     |
| blog-writing-tips  | 30 days ago | +5% QoQ        | Low      |

Configuration

Add the GSC MCP server to your Claude Code settings (~/.claude/settings.json):

{
  "mcpServers": {
    "google-search-console": {
      "command": "npx",
      "args": ["-y", "@anthropic/gsc-mcp-server"],
      "env": {
        "GSC_CREDENTIALS_PATH": "/path/to/credentials.json"
      }
    }
  }
}

Setup requirements:

  1. Google Cloud project with Search Console API enabled
  2. OAuth credentials or service account key
  3. Site verified in Google Search Console

Ahrefs MCP

What It Enables

Feature Without Ahrefs MCP With Ahrefs MCP
Backlink analysis Not available Referring domains, anchor text distribution
Keyword research WebSearch only Search volume, difficulty, SERP features
Competitor monitoring Manual WebSearch Automated gap analysis and tracking
Content gap analysis Not available Keywords competitors rank for that you don't
Domain Rating Not available Live DR tracking

Enhanced Workflows

/blog brief with Ahrefs data:

Content briefs can include precise keyword metrics:

Target Keywords
- Primary: "kubernetes monitoring" (2,400/mo, KD 45)
- Secondary: "k8s observability" (890/mo, KD 32)
- Question: "how to monitor kubernetes clusters" (720/mo, KD 28)

Competitor Content Gap
| Keyword                    | Competitor A | Competitor B | You  |
|---------------------------|-------------|-------------|------|
| kubernetes alerting setup  | #3          | #7          |:   |
| prometheus vs datadog      | #5          | #2          |:   |
| k8s monitoring best practices | #1       | #4          | #12  |

/blog strategy with Ahrefs data:

Strategy documents gain competitive intelligence with actual metrics:

Competitive Landscape
| Competitor      | DR  | Blog Posts | Avg Traffic/Post | Top Keywords |
|----------------|-----|-----------|-----------------|-------------|
| competitor-a.com | 72  | 340       | 2,100           | 890         |
| competitor-b.com | 65  | 180       | 1,400           | 520         |
| your-site.com    | 45  | 47        | 380             | 120         |

Opportunity: 234 keywords where competitors rank but you don't

Configuration

{
  "mcpServers": {
    "ahrefs": {
      "command": "npx",
      "args": ["-y", "@anthropic/ahrefs-mcp-server"],
      "env": {
        "AHREFS_API_KEY": "your-api-key"
      }
    }
  }
}

Setup requirements:

  1. Ahrefs account with API access (Standard plan or higher)
  2. API key from Ahrefs dashboard

Semrush MCP

What It Enables

Feature Without Semrush MCP With Semrush MCP
Keyword gap analysis Not available Side-by-side keyword overlap with competitors
Position tracking Not available Daily rank tracking for target keywords
Topic research WebSearch only Semrush Topic Research data
Content audit Quality-only scoring Quality + traffic + keyword data

Enhanced Workflows

/blog strategy with Semrush data:

Topic research backed by Semrush's keyword clustering:

Content Pillar: Kubernetes Monitoring
| Topic Cluster      | Keywords | Total Volume | Avg KD | Gap Score |
|-------------------|----------|-------------|--------|-----------|
| Setup & Config     | 34       | 12,400      | 38     | High      |
| Tools Comparison   | 22       | 8,900       | 52     | Medium    |
| Best Practices     | 18       | 6,200       | 41     | High      |
| Troubleshooting    | 45       | 15,100      | 29     | Low       |

Configuration

{
  "mcpServers": {
    "semrush": {
      "command": "npx",
      "args": ["-y", "@anthropic/semrush-mcp-server"],
      "env": {
        "SEMRUSH_API_KEY": "your-api-key"
      }
    }
  }
}

PageSpeed Insights MCP

What It Enables

Feature Without PSI MCP With PSI MCP
Core Web Vitals Not available LCP, FID, CLS, INP measurements
TTFB monitoring Not available Server response time (critical for AI crawlers)
Performance scoring Not available Lighthouse performance score
AI crawl readiness Manual check Automated TTFB < 200ms verification

Enhanced Workflows

/blog geo with PageSpeed data:

AI citation audits can include technical performance checks:

AI Crawler Readiness
| Metric   | Value  | Target  | Status |
|----------|--------|---------|--------|
| TTFB     | 145ms  | < 200ms | Pass   |
| LCP      | 2.1s   | < 2.5s  | Pass   |
| CLS      | 0.08   | < 0.1   | Pass   |
| JS-only? | No     | No      | Pass   |

Your pages are accessible to AI crawlers (GPTBot, ClaudeBot, PerplexityBot).

Configuration

{
  "mcpServers": {
    "pagespeed": {
      "command": "npx",
      "args": ["-y", "@anthropic/pagespeed-mcp-server"],
      "env": {
        "PAGESPEED_API_KEY": "your-google-api-key"
      }
    }
  }
}

Setup requirements:

  1. Google Cloud project with PageSpeed Insights API enabled
  2. API key from Google Cloud Console

How to Configure MCP Servers

MCP servers are configured in your Claude Code settings file. The location depends on your setup:

Settings File Location

Platform Path
Linux/macOS ~/.claude/settings.json
Windows %USERPROFILE%\.claude\settings.json

Adding an MCP Server

Edit settings.json to add MCP servers under the mcpServers key:

{
  "mcpServers": {
    "server-name": {
      "command": "npx",
      "args": ["-y", "package-name"],
      "env": {
        "API_KEY": "your-key"
      }
    }
  }
}

Verifying MCP Connection

After adding an MCP server:

  1. Restart Claude Code
  2. The MCP server should appear in available tools
  3. Test with a simple query related to the server's function

Environment Variables for API Keys

Never commit API keys to version control. Use environment variables:

# Add to ~/.bashrc or ~/.zshrc
export AHREFS_API_KEY="your-key"
export SEMRUSH_API_KEY="your-key"
export GSC_CREDENTIALS_PATH="/path/to/credentials.json"
export PAGESPEED_API_KEY="your-key"

Then reference them in settings:

{
  "mcpServers": {
    "ahrefs": {
      "command": "npx",
      "args": ["-y", "@anthropic/ahrefs-mcp-server"],
      "env": {
        "AHREFS_API_KEY": "${AHREFS_API_KEY}"
      }
    }
  }
}

Example Workflows

Content Decay Detection (GSC + Blog Audit)

Combine Google Search Console data with claude-blog's quality scoring to identify posts that need immediate attention:

1. /blog audit content/blog/        # Quality scores for all posts
2. GSC MCP provides traffic data    # QoQ traffic changes
3. Combined report identifies:
   - High quality but declining traffic  --> needs freshness update
   - Low quality and declining traffic   --> needs full rewrite
   - Low quality but stable traffic      --> optimize for AI citations
4. /blog calendar                    # Auto-prioritized update schedule

Competitor-Informed Content Strategy (Ahrefs + Strategy)

Use Ahrefs data to ground your content strategy in competitive intelligence:

1. /blog strategy "your-niche"      # Base strategy from analysis
2. Ahrefs MCP provides:
   - Competitor keyword rankings
   - Content gap keywords
   - Backlink opportunities
3. Strategy document includes:
   - Data-backed pillar topics
   - Specific keyword targets with volume/difficulty
   - Competitor weakness mapping
4. /blog calendar                    # Execution plan with keyword targets

Performance-Optimized GEO Audit (PSI + GEO)

Validate both content quality and technical readiness for AI crawlers:

1. /blog geo content/blog/post.mdx  # Content-level GEO audit
2. PageSpeed MCP provides:
   - TTFB measurement (must be < 200ms)
   - JavaScript rendering check
   - Core Web Vitals scores
3. Combined report covers:
   - Content optimization (answer-first, freshness, FAQ)
   - Technical optimization (TTFB, SSR, robots.txt)
   - AI crawler accessibility verification

Roadmap

Integration Status Priority
Nano Banana (Gemini) Available AI image generation for blog content
DataForSEO Available Recommended: covers SERP, keywords, backlinks, on-page, domain, content, AI optimization
Google Search Console Available (via /blog google gsc) First-party traffic/CTR data
Google Analytics (GA4) Available (via /blog google ga4) Organic traffic reports
WordPress REST API Future Low
Contentful / Sanity CMS Future Low

Community contributions for MCP server implementations are welcome. See CONTRIBUTING.md for guidelines.