YouTube Comments Scraper

Scrapes YouTube comments from channels or specific videos without an API key. It handles reply threading, incremental updates, and includes built-in state management for resumable runs.

Key Features

• Channel or Video Scraping: Provide a channel URL or a list of specific video IDs.
• No API Key: Scrapes directly without authentication or quotas.
• Incremental Mode: When enabled, only harvests new comments since the last run, drastically reducing cost for scheduled tasks.
• Complete Reply Threading: Captures nested replies with parent-child relationships.
• State Management: Saves checkpoints after each video, allowing runs to be resumed if interrupted.
• Flexible Configuration: Control limits per video, date ranges, concurrency, and proxy usage.

How to Use

Configure the actor run via input JSON. You must provide either a channelUrl or an array of videoIds.

Example Inputs

Scrape specific videos:

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

Scrape a channel:

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

Input Configuration

Output

Each comment is output as a structured JSON object. Example:

{
  "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"
}

Notes on Usage

• Pricing: You pay per comment successfully harvested. Using incrementalMode for scheduled runs minimizes cost by avoiding duplicate work.
• Proxies: For light scraping, datacenter proxies (default) are sufficient. For heavy, frequent, or concurrent runs, consider enabling useResidentialProxies: true (additional cost applies).
• Incremental Workflow: Set up a daily task with incrementalMode: true. The actor tracks state, so each run only processes and charges you for new comments.