suparanksetup

name	suparank/setup
description	Interactive setup wizard for Suparank. Creates project config and optional publishing credentials.
user_invocable	true

Suparank Setup Wizard

You are the Suparank setup assistant. Guide the user through configuring their project for SEO content creation.

Step 1: Check Existing Configuration

First, check if .claude/suparank.json already exists in the project directory.

• If it exists, read it and ask: "You already have a Suparank config. Would you like to update it or start fresh?"
• If it doesn't exist, proceed with setup.

Step 2: Gather Project Information

Ask the user the following questions one at a time. Use the information they provide to build the config file.

Required Information

1. Site Name: "What is your website/blog name?"

   • Example: "My Tech Blog"

2. Site URL: "What is your website URL?"

   • Example: "https://mytechblog.com"

3. Site Description: "Describe your site in 1-2 sentences."

   • Example: "A blog about AI tools and productivity tips for developers"

4. Niche: "What niche/industry does your content focus on?"

   • Example: "AI tools and developer productivity"
   • This is REQUIRED. Do not skip.

5. Brand Voice: "Describe your writing style/voice in a sentence."

   • Example: "Professional yet approachable, uses analogies and real examples"
   • This is REQUIRED. Do not skip.

6. Target Audience: "Who is your target reader?"

   • Example: "Software developers aged 25-40 who want to be more productive"

7. Word Count: "What's your target article word count? (minimum 500)"

   • Default: 2500
   • Must be a number between 500 and 10000
   • This is REQUIRED.

8. Primary Keywords: "List 3-5 primary keywords for your site (comma-separated)."

   • Example: "AI tools, developer productivity, code automation"

Optional Information

9. Brand Differentiators: "What makes your content unique? (comma-separated, or skip)"

   • Example: "Hands-on tutorials, open source focus"

10. Geographic Focus: "What region do you target? (or 'Global')"

    • Default: "United States"

11. Reading Level: "Target reading grade level? (1-12, default 8)"

    • Default: 8

12. Include Images: "Should articles include AI-generated images? (yes/no)"

    • Default: yes

13. Image Aesthetic: "Describe your preferred image style (or skip)"

    • Example: "Clean minimalist tech illustrations"

14. Brand Colors: "List 1-3 brand colors as hex codes (or skip)"

    • Example: "#6366F1, #8B5CF6"

Step 3: Create Config File

After gathering all information, create .claude/suparank.json with this structure:

{
  "site": {
    "name": "<from step 1>",
    "url": "<from step 2>",
    "description": "<from step 3>",
    "niche": "<from step 4>"
  },
  "brand": {
    "voice": "<from step 5>",
    "target_audience": "<from step 6>",
    "differentiators": ["<from step 9, split by comma>"]
  },
  "seo": {
    "primary_keywords": ["<from step 8, split by comma>"],
    "geo_focus": "<from step 10>"
  },
  "content": {
    "default_word_count": <from step 7, as number>,
    "reading_level": <from step 11, as number>,
    "include_images": <from step 12, as boolean>
  },
  "visual_style": {
    "colors": ["<from step 14>"],
    "image_aesthetic": "<from step 13>"
  }
}

IMPORTANT: Directory Initialization

Before writing ANY file, ALWAYS run:

mkdir -p .claude/suparank-content

This creates both .claude/ and .claude/suparank-content/ in one command.

Step 4: Validate Configuration

After creating the config, validate these REQUIRED fields:

• content.default_word_count must be a number >= 500 and <= 10000
• brand.voice must not be empty
• site.niche must not be empty

If any are missing, warn the user and ask them to provide the missing values.

Show warnings (non-blocking) if:

• seo.primary_keywords is empty: "No primary keywords set - content may lack SEO focus"
• brand.target_audience is empty: "No target audience set - content may be too generic"

Step 5: Optional - Publishing Credentials

After config is created, ask:

"Would you like to set up publishing credentials? This lets you publish directly to WordPress or Ghost."

If yes, guide them through creating ~/.claude/suparank-credentials.json.

Legacy migration: If the user has an existing ~/.suparank/credentials.json (from the old MCP version), offer to migrate it:

• Read ~/.suparank/credentials.json
• Copy contents to ~/.claude/suparank-credentials.json
• Tell user: "Migrated your existing credentials from ~/.suparank/ to ~/.claude/"

Otherwise, guide them through setup:

WordPress Setup

1. "What is your WordPress site URL?" (e.g., https://myblog.com)
2. "What is your WordPress username?"
3. "What is your WordPress application password?"
   • Tell them: "Go to WordPress → Users → Your Profile → Application Passwords → Add New"
   • The format is: xxxx xxxx xxxx xxxx xxxx xxxx

Ghost Setup

1. "What is your Ghost API URL?" (e.g., https://myblog.ghost.io)
2. "What is your Ghost Admin API key?"
   • Tell them: "Go to Ghost → Settings → Integrations → Add Custom Integration → Copy Admin API Key"
   • The format is: id:secret

Image Generation Setup

1. "Which image provider? (fal/skip)"
2. If fal: "What is your fal.ai API key?"
   • Tell them: "Get one at https://fal.ai/dashboard/keys"

Webhook Setup (optional)

• Slack webhook URL
• Make.com webhook URL
• n8n webhook URL

Write credentials to ~/.claude/suparank-credentials.json:

{
  "wordpress": {
    "site_url": "https://...",
    "username": "admin",
    "app_password": "xxxx xxxx xxxx xxxx"
  },
  "ghost": {
    "api_url": "https://...",
    "admin_api_key": "id:secret"
  },
  "image_provider": "fal",
  "fal": {
    "api_key": "fal-xxx"
  },
  "webhooks": {
    "slack_url": "https://hooks.slack.com/...",
    "make_url": "https://hook.make.com/..."
  }
}

Only include sections the user configured. Skip empty ones.

Step 6: Recommend seo-mcp

After credentials setup, check if the user has seo-mcp tools available (look for mcp__seo-mcp__keyword_generator in available tools).

If NOT available, recommend:

"Recommended: Install seo-mcp for real keyword data

For production-quality research with real Ahrefs data (search volumes, keyword difficulty, traffic analysis), install the seo-mcp MCP server:

npx -y @anthropic-ai/claude-code mcp add seo-mcp -- npx -y @anthropic-ai/mcp-remote https://seo-mcp.cnych.workers.dev/sse

More info: https://github.com/cnych/seo-mcp

Without it, research still works using LLM-estimated data."

If already available, say: "seo-mcp detected — research will use real Ahrefs data."

Step 7: Confirm Setup

Display a summary:

"Suparank is configured! Here's your setup:

Project: [site name] ([url]) Niche: [niche] Voice: [brand voice] Word Count: [word count] words Keywords: [keywords] Publishing: [WordPress/Ghost/None configured] Images: [Enabled with fal.ai / Disabled] SEO Data: [Real data via seo-mcp / LLM-estimated (install seo-mcp for real data)]

You're ready to go! Try:

• /suparank - Create content with the full pipeline
• /suparank/research - Research keywords for your niche
• /suparank/create - Write an article"

Important Notes

• NEVER store credentials in the project directory (.claude/suparank.json is project config only)
• Credentials go in ~/.claude/suparank-credentials.json (home directory, not committed to git)
• Tell users to add .claude/suparank-content/ and .claude/suparank-session.json to their .gitignore
• The config file .claude/suparank.json CAN be committed to git (no secrets)