Automate LinkedIn Engagement with GPT-4o Sentiment Analysis & Smart Comments

Description

LinkedIn Auto-Engagement AI Agent

Automatically find, analyze, and engage with relevant LinkedIn posts using AI-powered sentiment analysis and comment generation

📋 Template Overview

This workflow automates LinkedIn engagement by searching for relevant posts, analyzing their sentiment, generating contextual comments using AI, and posting them automatically. Perfect for DevOps professionals, cloud architects, and automation enthusiasts who want to maintain an active LinkedIn presence without spending hours manually commenting.

🎯 What This Workflow Does

Searches LinkedIn for posts from the last 24 hours based on your chosen keywords (Kubernetes, Azure, AWS, AI, Automation, SaaS, n8n)
Filters quality content - Only engages with posts over 350 characters that are original (not reposts)
Analyzes sentiment using AI to categorize posts into 7 types:
Informational/Educational Content
Happy Announcement
Sad/Negative Feelings
Hiring Posts
Not English Language
Neutral
None of the Above
Generates contextual comments using AI that are witty, professional, and add genuine value
Posts comments and adds reactions to the LinkedIn posts
Logs everything to Google Sheets for tracking and analysis
Sends Telegram notifications when the workflow completes

🔧 Prerequisites

Before you can use this workflow, you'll need:

Required Accounts & Credentials

n8n Account (Cloud or Self-hosted)
Unipile Account - For LinkedIn API access
Sign up at: https://unipile.com
Get your API key
Connect your LinkedIn account to Unipile
OpenRouter Account - For AI capabilities
Sign up at: https://openrouter.ai
Get your API key
Recommended model: openai/gpt-4o-mini (cost-effective)
Google Account - For Google Sheets logging
Telegram Account - For notifications (optional but recommended)

Technical Requirements

Basic understanding of n8n workflows
Ability to create and manage Google Sheets
Ability to create a Telegram bot (5 minutes, step-by-step below)

📦 Setup Instructions

Step 1: Import the Workflow

Copy the workflow JSON
In n8n, click + Add workflow
Click the three dots (...) in the top right
Select Import from File or Import from URL
Paste the JSON

Step 2: Set Up Google Sheets

Create a new Google Sheet with these exact column headers in Row 1:

| Post URL | Post Content | Name of Author | Author's Profile Headline | Profile URL of Author | ID | Sentiment of Post | Comment Generated by AI | Error |
|----------|--------------|----------------|---------------------------|----------------------|-------|-------------------|------------------------|--------|

Name the sheet tab: "Open - New n Random Niche Posts Comment"
Share the sheet with your Google account connected to n8n
Copy the Sheet ID from the URL (between /d/ and /edit)

Step 3: Configure Google Sheets Nodes

Update the Sheet ID in these 4 nodes:
Append row in sheet2
Update row in sheet9
Update row in sheet10
Update row in sheet11

For each node:
Click the node
Find the Document field
Either select your sheet from the dropdown OR paste your Sheet ID
Click the Refresh icon to load columns
Save

Step 4: Set Up Unipile (LinkedIn API)

Sign up at Unipile: https://unipile.com
Connect your LinkedIn account in the Unipile dashboard
Get your Account ID and API Key
Update these nodes with your credentials:
LinkedIn Search
Get Post from LinkedIn2
Comment On Post2
Leave Reaction on Post2

For each node:
Find the X-API-KEY header parameter
Paste your Unipile API key
Find the account_id parameter
Paste your Unipile account ID

Step 5: Set Up OpenRouter (AI)

Sign up at OpenRouter: https://openrouter.ai
Get your API key from the dashboard
Add credits to your account ($5-10 is plenty to start)
Configure these nodes:
OpenRouter Chat Model6
OpenRouter Chat Model7
OpenRouter Chat Model8

For each node:
Click the node
Click Credentials dropdown
Select Create New
Paste your OpenRouter API key
Save

Step 6: Set Up Telegram Notifications (Optional)

Create a Telegram Bot:
Open Telegram and search for @BotFather
Send /newbot
Choose a name (e.g., "LinkedIn Bot")
Choose a username (must end in 'bot')
Save the API Token BotFather gives you

Get Your Chat ID:
Start a chat with your bot
Send any message
Visit: https://api.telegram.org/botYOUR_TOKEN/getUpdates
Find your Chat ID in the JSON response

Configure the Telegram Node:
Open Send a text message2 node
Add your Bot Token in credentials
Replace the chatId with your Chat ID

Step 7: Customize Keywords (Optional)

Want to search for different topics? Update the LinkedIn Search URL:

Current keywords: Kubernetes, Azure, n8n, AWS, AI, Automation, SaaS

To change:
Open LinkedIn Search node
Find the url parameter under Body Parameters
Modify the keywords= section

Example for DevOps tools:
keywords=%22Terraform%22%20OR%20%22Docker%22%20OR%20%22Kubernetes%22%20OR%20%22Jenkins%22

Step 8: Test the Workflow

Click Execute Workflow (or use Manual Trigger)
Watch each node execute
Check for any errors (red nodes)
Verify data appears in your Google Sheet
Check your Telegram for the completion message

🎨 Customization Options

Modify AI Comment Style

Edit the AI Agent2 node prompt to change comment style:

You're an expert on [YOUR EXPERTISE].

Look at this post and drop in a thoughtful comment - {{ $json['Post Content'] }}

Guidelines:
Write like a [YOUR STYLE]
[YOUR RULES]
Maximum 2 sentences

Only export the response and nothing else.

Change Sentiment Categories

Edit the Sentiment Analysis2 node to add/remove categories based on your needs.

Adjust Filtering Criteria

Modify the If2 node to change:
Minimum character count (currently 350)
Post type filters
Date range filters

Schedule Automation

Add a Schedule Trigger node to run automatically:
Recommended: 2-3 times per day (morning, afternoon, evening)
Avoid running too frequently to prevent spam detection

📊 Workflow Nodes Explained

Core Processing Flow

LinkedIn Search - Fetches posts from LinkedIn via Unipile API
Code4 - Parses the API response into individual posts
Code5 - Filters posts to last 24 hours only
Basic LLM Chain2 - Counts characters in each post
If2 - Filters posts (350+ chars, original content only)
Append row in sheet2 - Logs qualifying posts to Google Sheets
Loop Over Items4 - Processes each post one by one

Analysis & Generation

Sentiment Analysis2 - AI categorizes post sentiment
AI Agent2 - Generates contextual comment
Get Post from LinkedIn2 - Fetches fresh post data
Comment On Post2 - Posts the AI-generated comment
Leave Reaction on Post2 - Adds a like to the post

Logging & Notifications

Update row in sheet9/10/11 - Updates Google Sheet with results
Wait nodes - Adds delays to avoid rate limiting
Send a text message2 - Telegram notification on completion

⚙️ Advanced Features

Rate Limiting

The workflow includes 1-minute wait times between actions to avoid LinkedIn's spam detection. You can adjust these in the Wait nodes.

Error Handling

Failed comments are logged in the "Error" column
Retry logic built into update nodes (5 retries, 5-second intervals)
Errors don't stop the workflow from processing other posts

Multi-Niche Support

To engage with multiple niches:
Duplicate the LinkedIn Search node
Change keywords in each duplicate
Add a Merge node to combine results
Connect to the rest of the workflow

🚀 Usage Tips

Best Practices

Start slow**: Run manually 2-3 times to ensure everything works
Monitor results**: Check your Google Sheet daily
Refine prompts**: Adjust AI prompts based on comment quality
Vary timing**: Schedule runs at different times of day
Stay authentic**: Review and approve comments if needed

What to Avoid

❌ Running too frequently (max 3-4 times per day)
❌ Using generic keywords that return spam
❌ Commenting on your own posts
❌ Using the same comment style repeatedly
❌ Engaging with low-quality content

LinkedIn Best Practices

Focus on adding genuine value
Engage with diverse content (not just one niche)
Avoid hashtags in comments
Keep comments concise (2 sentences max)
Build relationships, not just visibility

📈 Performance & Costs

Expected Results

Posts found**: 20-100 per run (depending on keywords)
Posts filtered**: 10-30 qualifying posts
Comments posted**: 5-15 per run (after sentiment filtering)
Time saved**: 1-2 hours of manual engagement per day

API Costs (Approximate)

Unipile**: ~$29/month for LinkedIn API access
OpenRouter**: Free
n8n**: Free (self-hosted)Subscript
Google Sheets**: Free
Telegram**: Free

Total estimated cost: $30-50/month for full automation

🐛 Troubleshooting

Common Issues

Issue: "Column names were updated" error
Solution**: Refresh columns in Google Sheets nodes

Issue: LinkedIn API returns 403 error
Solution**: Check Unipile account status and API limits

Issue: AI comments are generic or low quality
Solution**: Refine the AI Agent2 prompt with more specific guidelines

Issue: No posts found
Solution**: Check keyword relevance and time range settings

Issue: Workflow hangs on Loop Over Items
Solution**: Check wait times and ensure all nodes are properly connected

Getting Help

Join the n8n community: https://community.n8n.io
Check n8n docs: https://docs.n8n.io
Unipile documentation: https://docs.unipile.com

🔐 Security & Privacy

Data Protection

All LinkedIn credentials are handled by Unipile (OAuth)
API keys are stored securely in n8n credentials
Google Sheets data is private to your account
No third-party data sharing

Compliance

Respects LinkedIn's rate limits
Follows LinkedIn's Terms of Service
Only engages with public posts
No scraping of private data

Ethical Use

This workflow is designed for:
✅ Authentic professional networking
✅ Adding value to discussions
✅ Saving time on routine engagement

Please avoid:
❌ Spamming or aggressive self-promotion
❌ Misleading or deceptive comments
❌ Violating LinkedIn's community guidelines

📝 Changelog & Updates

Version 1.0 (Current)
Initial release
Core functionality: search, filter, analyze, comment
Google Sheets logging
Telegram notifications
Multi-keyword support

Planned Features
Multi-account support
Advanced scheduling options
Comment approval workflow
Analytics dashboard
A/B testing for comment styles

👨‍💻 About the Creator

This workflow was created to solve the challenge of maintaining consistent LinkedIn engagement while focusing on actual work. After spending too many hours manually commenting, I built this automation to handle the repetitive parts while keeping comments authentic and valuable.

📄 License

This workflow template is free to use and modify for personal and commercial purposes. Attribution is appreciated but not required.

🤝 Contributing

Found a bug? Have an improvement idea?

Feel free to:
Share your modifications
Report issues in the n8n community
Suggest new features
Create your own variations

🙏 Acknowledgments

Built with:
n8n - The workflow automation platform
Unipile - LinkedIn API integration
OpenRouter - AI model access
Google Sheets - Data logging
Telegram - Notifications

Ready to automate your LinkedIn engagement? Import this workflow and start building your professional network on autopilot!

Last updated: October 2025