YouTube Comments Harvester

Comprehensive YouTube comments scraper with channel-wide enumeration, incremental scraping, complete reply threading, and de-duplication. No API key required.

Features

• 🎯 Channel-wide harvesting: Auto-discover and process all videos from a channel
• 🔄 Incremental mode: Only harvest NEW comments since last run (save costs)
• 💬 Complete reply threading: Extract all nested replies with parent-child relationships
• 🚫 No API key required: Scrape without quotas or authentication
• 💾 State management: Resumable runs with checkpoints after each video
• 🎛️ Flexible input: Process by channel URL or specific video IDs

Use Cases

• Brand monitoring: Track comments on your videos or competitors' videos
• Sentiment analysis: Analyze audience reactions and opinions
• Competitor research: Understand what resonates with competitor audiences
• Content strategy: Identify trending topics and popular discussions
• AI/ML datasets: Collect training data for NLP models
• Scheduled monitoring: Run daily/weekly to track new comments

Input Configuration

Required Fields

Either channelUrl OR videoIds must be provided.

Example Inputs

Process specific videos:

{
  "videoIds": ["dQw4w9WgXcQ", "jNQXAC9IVRw"],
  "maxCommentsPerVideo": 100,
  "includeReplies": true
}

Process channel videos:

{
  "channelUrl": "@veritasium",
  "maxVideos": 10,
  "maxCommentsPerVideo": 50,
  "since": "2024-01-01",
  "incrementalMode": true
}

Input Fields

Field	Type	Default	Description
channelUrl	string	-	YouTube channel URL (e.g., @veritasium or https://youtube.com/@veritasium)
videoIds	array	[]	Specific video IDs to process (alternative to channelUrl)
maxVideos	integer	10	Maximum videos to process from channel (1-1000)
maxCommentsPerVideo	integer	100	Maximum comments per video (1-10000)
includeReplies	boolean	true	Include nested comment replies
since	string	-	Only process videos after this date (ISO 8601: YYYY-MM-DD)
incrementalMode	boolean	false	Only harvest NEW comments since last run
useResidentialProxies	boolean	false	Enable residential proxies (recommended for heavy scraping, higher cost)
maxConcurrency	integer	2	Videos to process simultaneously (1-5)

Output Schema

Each comment is saved with the following structure:

{
  "commentId": "UgxKREPabcdefghijk",
  "videoId": "dQw4w9WgXcQ",
  "videoUrl": "https://youtube.com/watch?v=dQw4w9WgXcQ",
  "videoTitle": "Rick Astley - Never Gonna Give You Up",
  "videoPublishedAt": "2 years ago",
  "text": "This song never gets old!",
  "authorName": "John Doe",
  "authorChannelId": "UCabcdefghijk123456789",
  "likeCount": 42,
  "replyCount": 3,
  "publishedAt": "2 days ago",
  "isReply": false,
  "parentCommentId": null,
  "scrapedAt": "2025-10-19T12:00:00.000Z"
}

Incremental Scraping

Enable incrementalMode: true for scheduled runs to avoid re-processing comments:

• State management: Tracks processed videos and seen comment IDs
• De-duplication: Only saves NEW comments since last run
• Cost savings: Only pay for new comments (not duplicates)
• Resumable: Checkpoints after each video

Example workflow:

1. Run daily with incrementalMode: true
2. Only new comments are harvested
3. Costs 99% less than full re-scraping

Pricing

• Model: Pay-per-event ($0.001 per comment)
• Cost: $1.00 per 1,000 comments
• Incremental advantage: Only pay for NEW comments

Example costs:

• 10 videos × 100 comments = 1,000 comments = $1.00
• Channel with 50 videos × 50 comments = 2,500 comments = $2.50
• Daily monitoring (50 new comments/day) = $0.05/day = $1.50/month

Proxy Usage

When proxies are needed:

• Heavy scraping (>1,000 comments)
• Multiple concurrent runs
• Frequent runs (hourly/daily)

Cost comparison:

• Datacenter proxies (default): Included in compute
• Residential proxies (useResidentialProxies: true): ~$2-5 per 1,000 comments (additional)

Limitations

• YouTube may rate-limit aggressive scraping (use proxies if needed)
• Relative timestamps ("2 days ago") not exact ISO dates
• Initial page load only captures first ~20-100 comments (scrolling loads more)
• Continuation tokens for deep reply pagination not yet implemented

Technical Details

Scraping method:

• Playwright browser automation
• Extracts ytInitialData from page source (YouTube's embedded JSON)
• Scrolls to load more comments
• Parses comment threads with parent-child relationships

No API required:

• Works without YouTube Data API key
• No quota limits
• Fully autonomous

🔗 Integration with Automation Tools

n8n Workflow Example

Daily Comment Monitoring:

Schedule (9am daily)
  ↓
Apify: Run YouTube Comments Harvester (incrementalMode: true)
  ↓
Filter: NEW comments only
  ↓
OpenAI: Sentiment analysis
  ↓
Airtable: Save to "Comments" base
  ↓
Slack: Alert on negative sentiment

Zapier Workflow Example

Brand Mention Tracking:

1. Schedule: Every 6 hours
2. Run Apify actor (YouTube Comments Harvester)
3. Filter: text contains "your-brand-name"
4. Google Sheets: Append to tracking sheet
5. Gmail: Send digest email

Make.com Workflow Example

Competitor Analysis:

1. YouTube Harvester: Scrape competitor channels
2. Iterator: Process each comment
3. HTTP: Keyword extraction via ChatGPT
4. Airtable: Aggregate trending topics
5. Notion: Update research dashboard

❓ FAQ

Do I need a YouTube API key?

No! This scraper extracts data directly from YouTube pages using browser automation. No API quotas or authentication required.

Can I use this with n8n, Zapier, or Make?

Yes! Export data to CSV/JSON or use Apify's API integration with automation tools. The flat output structure works perfectly with n8n, Zapier, and Make workflows.

How does incremental mode save costs?

Incremental mode tracks which videos and comments you've already scraped. On subsequent runs, it only harvests NEW comments, so you're not re-paying for the same data. For daily monitoring, this can reduce costs by 99%.

Can I scrape comment replies?

Yes! Set includeReplies: true and configure commentDepth (0-5 levels). The scraper extracts nested replies with parent-child relationships preserved.

How many comments can I scrape per video?

Technically unlimited with maxCommentsPerVideo: 0, but YouTube typically loads the first ~100-200 comments by default. Scrolling loads more, but extremely long comment threads may take time.

Can I scrape private or unlisted videos?

No, only public YouTube videos and comments are accessible. Private/unlisted videos require authentication and aren't supported.

How accurate are the timestamps?

YouTube provides relative timestamps ("2 days ago") rather than exact ISO dates. The scraper preserves these as-is. For exact timestamps, you'd need YouTube Data API access.

When should I use residential proxies?

For heavy scraping (>1,000 comments per run), frequent runs (hourly), or if you're getting blocked. Otherwise, the default datacenter proxies work fine.

How do I schedule recurring runs?

Use Apify's built-in scheduler or integrate with n8n/Zapier. For brand monitoring, we recommend daily runs with incrementalMode: true.

What format is the comment data?

Flat JSON with all fields at the root level - perfect for importing to Google Sheets, databases, or automation tools. Export as CSV, JSON, or Excel.

---

🔗 Explore More of Our Actors

💬 Social Media & Community

Actor	Description
Reddit Scraper Pro	Monitor subreddits and track keywords with sentiment analysis
Discord Scraper Pro	Extract Discord messages and chat history for community insights
YouTube Contact Scraper	Extract YouTube channel contact information for outreach
YouTube Shorts Scraper	Scrape YouTube Shorts for viral content research

🛒 E-commerce & Brand Monitoring

Actor	Description
Shopify Scraper Pro	Extract Shopify product data for competitor and brand monitoring
Amazon Reviews Scraper	Extract Amazon customer reviews for sentiment analysis

🏢 Business Intelligence

Actor	Description
Indeed Salary Analyzer	Get salary data for compensation benchmarking and market research
Crunchbase Scraper	Extract company data and funding information for business intelligence

---

🏷️ SEO Keywords

YouTube comment scraper, YouTube data extraction, scrape YouTube comments, YouTube API alternative, extract YouTube comments, YouTube comment harvester, YouTube channel scraper, YouTube automation, YouTube comment analysis, YouTube sentiment analysis, YouTube brand monitoring, YouTube competitor analysis, how to scrape YouTube, YouTube data mining, YouTube comment export, n8n YouTube integration, Zapier YouTube scraper, Make YouTube automation, YouTube research tool, YouTube engagement data

Support

For issues or questions:

• Email: kontakt@barrierefix.de
• GitHub: Report issues in actor repository

---

Note: This actor scrapes public YouTube data. Ensure your use complies with YouTube's Terms of Service and applicable laws.