Perplexity Research Pipeline

A powerful Claude Code skill that automates research-to-content workflows using Perplexity AI and browser automation. Perfect for SEO content creators, researchers, and anyone who needs systematic, deep research before writing.

🚀 Features

• Automated Perplexity Research: Uses browser automation to run deep research with Twitter/Reddit sentiment analysis
• Research-to-Prompt Pipeline: Automatically converts research findings into optimized writing prompts
• CDP Browser Control: Leverages Chrome DevTools Protocol for reliable browser operations
• Smart Content Generation: Integrates with ChatGPT for high-quality, SEO-optimized content
• Batch Processing: Supports multi-keyword workflows with automatic progression
• Project Organization: Auto-creates project folders with structured file management

📋 Requirements

• Node.js 22+ - Required for URL encoding and CDP operations
• Chrome with Remote Debugging - Enable at chrome://inspect/#remote-debugging
• Claude Code - With web-access skill installed
• CDP Proxy - Started via web-access skill

🔧 Installation

Step 1: Install Dependencies

# Verify Node.js installation
node --version  # Should be v22+

# Enable Chrome Remote Debugging
# 1. Open chrome://inspect/#remote-debugging
# 2. Enable "Allow remote debugging for this browser instance"
# 3. Restart Chrome

Step 2: Install Web-Access Skill

# Install web-access skill (if not already installed)
# This skill depends on web-access for CDP browser control

Step 3: Install This Skill

# Clone to Claude Code skills directory
cd ~/.claude/skills
git clone https://github.com/jiemao777/perplexity-research-pipeline.git

# Or manually copy the folder
# Copy the entire folder to ~/.claude/skills/

📖 Usage

Basic Research Mode

Simply ask Claude Code to use Perplexity for research:

"Use Perplexity to research the latest trends in electric vehicle batteries"

Research-to-Prompt Pipeline (SEO Content)

For content creation workflows:

"Create a research-to-prompt pipeline for 'bulk feed additives supplier China' with cluster keywords:
1. import bulk feed additives from China
2. feed additives manufacturing process
3. bulk feed additives pricing and MOQ
4. global feed additives markets
5. FAQ for international buyers"

Batch Processing

"Execute all 5 cluster keywords for the amino acid project using Perplexity research"

🔄 Workflow

This skill follows a systematic 3-step process:

Step 1: Perplexity Research

• Opens Perplexity in automated browser
• Runs deep research with your custom prompt
• Includes Twitter/Reddit sentiment analysis
• Extracts comprehensive findings

Step 2: Research File Creation

• Saves cleaned research as Markdown
• Structures content with sections:
  • Topic and context
  • Key findings with data
  • Social media sentiment
  • Sources cited

Step 3: Final Prompt Assembly

• Combines research with content-generation template
• Fills in keywords, internal links, image embeds
• Produces ChatGPT-ready prompt file
• Stops for manual review (you control final generation)

📁 Project Structure

your-project-folder/
├── keyword-1-research.md
├── keyword-1-final-prompt.md
├── keyword-2-research.md
├── keyword-2-final-prompt.md
├── progress.txt
└── pillar-research.md
    └── pillar-final-prompt.md

⚙️ Configuration

Content-Generation Template

The skill uses a fixed template structure (defined in skill.md). Key sections:

1. Core Persona - Expert SEO + NLP writer identity
2. Writing Style - Simple, clear, active voice
3. Article Structure - Key Takeaways + 10-15 H2 headings + FAQ
4. SEO Requirements - Internal linking, image embedding
5. Prohibitions - Avoids AI/marketing fluff language

Customization Points

You can customize per article:

• <information> - Research content from Perplexity
• <keyword> - Target keywords (3-8 related keywords)
• <internal links> - Full URL list for internal linking
• <image embeds> - Image library with structured metadata

🎯 Best Practices

1. Research Prompt Design

For Cluster Pages:

I am writing an article for www.yoursite.com. I want you to do some research on the blog post topic cluster of "[CLUSTER KEYWORD]" for the pillar page "[PILLAR KEYWORD]" but the context of the business into account and including data. Also what are the feedbacks from twitter users and reddit.com?

For Pillar Pages:

I am writing an article for www.yoursite.com. I want you to do some research on the blog post pillar page of "[PILLAR KEYWORD]" and include the Topic Cluster "[CLUSTER 1]" "[CLUSTER 2]" "[CLUSTER 3]" "[CLUSTER 4]" "[CLUSTER 5]" but the context of the business into account and including data. Also what are the feedbacks from twitter users and reddit.com?

2. File Naming

Use descriptive slugs derived from keywords:

• l-lysine-feed-grade-research.md
• dl-methionine-feed-grade-final-prompt.md
• amino-acid-supplier-lysine-methionine-research.md

3. Content Sequencing

Recommended order for pillar + cluster sets:

1. Cluster page 1
2. Cluster page 2
3. Cluster page 3
4. Cluster page 4
5. Cluster page 5
6. Pillar page (uses all 5 clusters as context)

🔍 Troubleshooting

CDP Proxy Not Running

bash ~/.claude/skills/web-access/scripts/check-deps.sh

Chrome Not Connected

1. Open chrome://inspect/#remote-debugging
2. Enable "Allow remote debugging for this browser instance"
3. Restart Chrome

Cloudflare Blocking

• Perplexity may trigger Cloudflare challenges
• Pause and complete verification manually in browser
• Resume after verification

Target Lost/Navigation Failed

# List current tabs
curl -s http://localhost:3456/targets

# Re-create Perplexity tab
curl -s "http://localhost:3456/new?url=https://www.perplexity.ai/search?q=<ENCODED_PROMPT>"

📊 Example Output

Research File Structure

# Research: Your Topic Here

## Topic and Context
**Cluster Keyword:** your keyword here
**Pillar Keyword:** main pillar keyword
**Business Context:** www.yoursite.com

## Key Findings

### 1. Main Point
- Data point 1
- Data point 2

### 2. Another Point
- Supporting evidence

## Sources Referenced
1. Source A
2. Source B

Final Prompt Structure

1) Core Persona
Identity: Expert SEO + NLP content writer

2) Writing Style & Tone
- Use simple words
- Active voice
- Clear and concise

3) Article Structure
A) Key Takeaways
B) 10-15 H2 headings
C) FAQ section

4) SEO Requirements
- Internal linking strategy
- Image embedding rules

7) Inputs
<information>
[Research content here]
</information>

<keyword>
[Target keywords]
</keyword>

<internal links>
[URL list]
</internal links>

<image embeds>
[Image markdown]
</image embeds>

🤝 Contributing

Contributions welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Submit a pull request

📝 License

This skill is provided as-is for use with Claude Code. Feel free to adapt and modify for your needs.

🔗 Links

• Repository: https://github.com/jiemao777/perplexity-research-pipeline
• Dependencies: web-access skill
• Claude Code: https://claude.ai/code

💡 Use Cases

• SEO Content Creation: Systematic keyword research + content generation
• Market Research: Deep dives with social sentiment analysis
• Academic Research: Comprehensive source gathering
• Competitor Analysis: Structured research workflows
• Product Research: Feature comparisons and user feedback

🎓 Advanced Features

Batch Override Mode

For processing multiple keywords without manual URL callbacks:

"Finish all remaining keywords in one batch without waiting for published URLs"

This mode:

• Generates predicted URLs from slugs
• Updates internal links immediately
• Processes entire batch continuously

Custom Image Libraries

Maintain master image files with structured format:

# Image Embeds - Pillar Name

## Products
URL: https://example.com/product, Image URL: https://example.com/image.jpg, Title: Product Name

## Packaging
URL: https://example.com/packaging, Image URL: https://example.com/package.jpg, Title: Packaging Options

Converts to markdown automatically:

[![Product Name](https://example.com/image.jpg)](https://example.com/product)

---

Made with ❤️ for researchers and content creators

Version: 1.0.0 Last Updated: 2026-03-25