💬 YouTube Comments Scraper & Pain Point Analyzer

Get every comment, and find out what people actually complain about. Paste video links or just type a keyword like iphone 17 problems or notion review. The Actor reads the comments (with replies, likes, pinned and creator-hearted flags), labels each one as a complaint, request, question or praise, and builds an insights report with the top pain points, what people ask for and the most-liked complaints and questions.

• ✅ Comments + replies with likes, reply count, verified badge, pinned, ❤️ by creator, approximate date
• ✅ Keyword search: no links needed. Research a product, a competitor or a niche across the top videos in one run
• ✅ Every comment labelled: complaint, request, question, praise, other (English, Chinese, Spanish, Portuguese)
• ✅ Pain points and requests: recurring phrases in complaints and requests, weighted by likes, with example comments. Topic words (the product name, the video title) are filtered out automatically
• ✅ Top / Newest sort and last N days filter: cheap daily monitoring of new comments
• ✅ HTML insights report + clean JSON / CSV / Excel
• ✅ No API key, no login, no quota. $0.50 per 1,000 comments

🎯 Who is it for?

• Product managers and founders: what users hate and wish for, straight from review videos about your product or your competitors.
• YouTubers and content teams: the questions your audience keeps asking = your next videos.
• Marketers and agencies: real customer language for ads, landing pages and FAQs.
• Researchers and analysts: labelled comment datasets for sentiment and topic studies.

🚀 Quick start

1. Paste one or more video URLs, or type search keywords (e.g. iphone 17 problems).
2. Keep Pain point analysis on.
3. Click Start. Open the REPORT in the Storage tab for the insights, or download all comments as JSON, CSV or Excel.

⬇️ Input examples

What people complain about in iPhone 17 review videos:

{ "searchQueries": ["iphone 17 review problems"], "videosPerSearch": 5, "maxCommentsPerVideo": 300 }

Every comment and reply on one video:

{ "videoUrls": ["https://www.youtube.com/watch?v=MFgpi8IvMB8"], "maxCommentsPerVideo": 5000, "includeReplies": true, "maxRepliesPerComment": 50 }

Daily monitoring of new comments on your videos (schedule it):

{ "videoUrls": ["https://youtu.be/VIDEO_ID_1", "https://youtu.be/VIDEO_ID_2"], "sort": "newest", "onlyNewerThanDays": 1 }

⬆️ Output

One row per comment (replies are their own rows with parentId):

{
  "videoId": "MFgpi8IvMB8",
  "videoTitle": "iPhone 17 - Long Term, REAL WORLD Review!",
  "channel": "Arran Brown",
  "commentId": "UgzXvO...",
  "text": "Saw a lot of threads in reddit regarding the battery drain issues and heating 😐",
  "label": "complaint",
  "likes": 2,
  "replyCount": 1,
  "authorIsVerified": false,
  "authorIsChannelOwner": false,
  "isPinned": false,
  "isHeartedByCreator": false,
  "publishedTimeText": "3 months ago",
  "publishedAt": "2026-06-24T10:00:00.000Z",
  "parentId": null,
  "videoUrl": "https://www.youtube.com/watch?v=MFgpi8IvMB8&lc=UgzXvO..."
}

The key-value store has OUTPUT (the insights as JSON) and REPORT (HTML):

{
  "comments": 452,
  "mix": { "complaint": { "count": 17, "share": 0.038 }, "question": { "count": 37, "share": 0.082 }, "praise": { "count": 47, "share": 0.104 } },
  "painPoints": [
    { "theme": "battery drain", "comments": 3, "totalLikes": 2, "examples": [{ "text": "Saw a lot of threads in reddit regarding the battery drain issues and heating 😐", "likes": 2 }] },
    { "theme": "heating", "comments": 3, "totalLikes": 3 }
  ],
  "featureRequests": [],
  "topComplaints": [],
  "topQuestions": []
}

publishedAt is approximate: YouTube only shows "3 months ago".

💵 Pricing

Event	Price
Comment or reply returned	$0.0005 ($0.50 per 1,000)
Insights report (labels, pain points, requests), once per run with 20+ comments	$0.10

Example: 1,000 comments from 5 review videos + the insights report = $0.60. Turn the analysis off and you pay for comments only.

🤖 Use it from AI agents (MCP)

Add this URL as an MCP server in Claude, Cursor or any MCP client and ask "What do people complain about in Notion review videos?":

https://mcp.apify.com?tools=lsso/youtube-comments-pain-points

❓ FAQ

Do I need a YouTube API key? No. The Actor reads comments the way the YouTube website does. No key, no quota, no login.

How is a comment labelled? Multilingual keyword rules with negation handling ("no issues at all" is praise, not a complaint). They are fast, free and consistent. A comment needs a clear negative signal to count as a complaint.

What if comments are turned off? The video is listed in OUTPUT with status COMMENTS_DISABLED and nothing is charged for it.

Can I get all comments of a huge video? Yes, up to 20,000 top-level comments per video, plus replies. Large runs take a few minutes.

Is it legal? What about privacy? The Actor reads only public comments that anyone can see without logging in, and it does not output usernames or channel IDs of commenters. Use the results for research and analysis, not to contact or track individual people, and respect YouTube's Terms of Service and privacy laws such as GDPR.

📝 Changelog

• 1.2 (2026-09-25): Commenter usernames and channel IDs are no longer included in the output (privacy); the proxy is now opt-in only, so you never pay for proxy traffic you did not ask for; if YouTube rate-limits the run, it stops cleanly and you are charged only for comments delivered; the insights report is created only when there are at least 20 comments.

• 1.1 (2026-09-24): "Newest first" sorting works again after a YouTube layout change (daily monitoring with last N days returns the new comments); replies stop at your max-cost limit; large runs analyse the 5,000 most-liked comments to stay within memory; "no proxy" really means no proxy; videos with comments turned off no longer fail the run.

• 1.0 (2026-09-24): First release: comments and replies, keyword search, Top / Newest, last-N-days filter, complaint / request / question / praise labels, pain point and request themes, HTML insights report.