YouTube Video API

The YouTube Video API returns structured details for one video, including its channel, description, chapters, related and end-screen videos, comments, and pagination tokens. It does not require JavaScript rendering or add a render fee.

Key Features

• Machine-readable publication time: published_at is an RFC 3339 timestamp in UTC, suitable for sorting and video-age calculations.
• Stable channel identifiers: channel.channel_id and comments[].author.channel_id remain stable when a channel changes its display name or handle.
• Exact numeric counts: extracted_views, extracted_likes, and extracted_subscribers complement YouTube's formatted display strings.
• Complete video metadata: Retrieve the description and links, chapters, related videos, end-screen videos, and transcript availability.
• Comments and replies: Use continuation tokens to load comments, sort by newest or top, and fetch replies.
• Localization: Control display strings with hl and influence related-video ranking with gl.

Endpoint

GET https://api.scrape.do/plugin/google/youtube/video

Request Parameters

Required

Localization

Pagination

The endpoint detects the token type automatically and returns the corresponding collection.

Example Requests

Video Details

curl "https://api.scrape.do/plugin/google/youtube/video?v=dQw4w9WgXcQ&token=$TOKEN"

Localized Details

curl "https://api.scrape.do/plugin/google/youtube/video?v=dQw4w9WgXcQ&hl=fr&gl=fr&token=$TOKEN"

More Related Videos

Pass related_videos_next_page_token from the previous response:

curl "https://api.scrape.do/plugin/google/youtube/video?v=dQw4w9WgXcQ&next_page_token=CBQSDRIL...&token=$TOKEN"

Comments or Replies

Pass comments_next_page_token, a token from comments_sorting_token, or a comment's replies_next_page_token:

curl "https://api.scrape.do/plugin/google/youtube/video?v=dQw4w9WgXcQ&next_page_token=Eg0SC2RR...&token=$TOKEN"

Response Modes

The endpoint uses one response schema with three modes. Fields that do not apply to the current mode are omitted.

Video Details Response

{
  "search_parameters": {
    "engine": "google_youtube_video",
    "v": "dQw4w9WgXcQ",
    "gl": "us",
    "hl": "en"
  },
  "title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
  "thumbnail": "https://i.ytimg.com/vi_webp/dQw4w9WgXcQ/maxresdefault.webp",
  "channel": {
    "name": "Rick Astley",
    "channel_id": "UCuAXFkgsw1L7xaCfnd5JJOw",
    "link": "https://www.youtube.com/channel/UCuAXFkgsw1L7xaCfnd5JJOw",
    "thumbnail": "https://yt3.ggpht.com/...",
    "subscribers": "4.51M subscribers",
    "extracted_subscribers": 4510000,
    "verified": true
  },
  "views": "1,790,053,431 views",
  "extracted_views": 1790053431,
  "likes": "19M",
  "extracted_likes": 19222757,
  "published_date": "Oct 24, 2009",
  "published_at": "2009-10-25T06:57:33Z",
  "description": { "content": "The official video...", "links": [] },
  "chapters": [],
  "related_videos": [],
  "related_videos_next_page_token": "CBQSDRILZFF3NHc5V2dYY1E...",
  "end_screen_videos": [],
  "comments_next_page_token": "Eg0SC2RRdzR3OVdnWGNR...",
  "comments_sorting_token": [],
  "transcript": { "token": "CgtkUXc0dzlXZ1hjUQ..." }
}

Publication Fields

Use published_at for chronological sorting, precise video age, and rate-per-hour calculations. Use published_date when you need the text shown for the requested locale. On rare videos where YouTube does not provide an exact timestamp, published_at is omitted while published_date can still be present.

channel

Use channel_id as the durable identifier instead of parsing link, because channel handles and display names can change.

Views and Likes

description

{
  "content": "The official video for Never Gonna Give You Up...",
  "links": [
    {
      "start_index": 180,
      "length": 16,
      "text": "#rickastleynever",
      "url": "https://www.youtube.com/hashtag/rickastleynever"
    }
  ]
}

links[].start_index and links[].length use UTF-16 code units, matching JavaScript string indexing. External targets use YouTube's redirect URL when that is what appears on the page.

chapters

Present only when the creator has defined chapters.

[
  {
    "title": "Introduction",
    "thumbnail": "https://i.ytimg.com/vi/...",
    "time_start": 0
  },
  {
    "title": "Variables",
    "thumbnail": "https://i.ytimg.com/vi/...",
    "time_start": 208
  }
]

time_start is the chapter's start offset in seconds.

related_videos

The initial response usually includes about 20 sidebar suggestions. Use related_videos_next_page_token to retrieve more.

[
  {
    "video_id": "u_cLVh73u_o",
    "link": "https://www.youtube.com/watch?v=u_cLVh73u_o",
    "thumbnail": {
      "static": "https://i.ytimg.com/vi/u_cLVh73u_o/hqdefault.jpg",
      "rich": "https://i.ytimg.com/an_webp/u_cLVh73u_o/mqdefault_6s.webp"
    },
    "title": "80'S MUSIC ON THE ROAD",
    "published_date": "6 days ago",
    "views": "421K views",
    "extracted_views": 421000,
    "length": "2:16:49",
    "channel": {
      "name": "Every Song Is A Scar",
      "link": "https://www.youtube.com/@everysongisascar",
      "verified": true
    }
  }
]

thumbnail.rich is an animated preview and is omitted when unavailable. A live suggestion has live: true and a watcher count, so views, extracted_views, and published_date are omitted. Shorts shelves between suggestions are not included.

end_screen_videos

End-screen cards include fields such as video_id, link, title, thumbnail, views, extracted_views, and length. YouTube does not display a publication date or channel on these cards, so those fields are absent.

comments_sorting_token

[
  { "title": "Top", "token": "Eg0SC2RRdzR3OVdnWGNR..." },
  { "title": "Newest", "token": "Eg0SC2RRdzR3OVdnWGNR..." }
]

Pass either token as next_page_token to load comments in that order. The initial comments_next_page_token uses the default Top ordering.

transcript

{ "token": "CgtkUXc0dzlXZ1hjUQ..." }

This opaque token indicates that the video has a transcript. It is reserved for future use; this endpoint does not return transcript text.

Comments Response

Comment pages contain 20 items per batch. Reply pages contain 10 items per batch.

{
  "search_parameters": {
    "engine": "google_youtube_video",
    "v": "dQw4w9WgXcQ",
    "gl": "us",
    "hl": "en",
    "next_page_token": "Eg0SC2RR..."
  },
  "comments": [
    {
      "comment_id": "UgzB3ZO9YOWY0PJqmZR4AaABAg",
      "content": "can confirm: he never gave us up",
      "published_date": "1 year ago",
      "likes": "265K",
      "extracted_likes": 265000,
      "replies_count": 961,
      "pinned": true,
      "author": {
        "name": "@YouTube",
        "link": "https://www.youtube.com/@YouTube",
        "channel_id": "UCBR8-60-B28hp2BmDPdntcQ",
        "thumbnail": "https://yt3.ggpht.com/...",
        "verified": true
      },
      "replies_next_page_token": "Eg0SC2RRdzR3OVdnWGNR..."
    }
  ],
  "comments_next_page_token": "Eg0SC2RRdzR3OVdnWGNR..."
}

Reply batches use the same comments item shape. They omit replies_next_page_token and use comments_next_page_token for additional replies in the same thread.

Pagination Guidelines

• Treat every continuation token as opaque and pass it back unchanged.
• Tokens belong to a specific video, collection, and position. Do not reuse them across videos.
• Stop when the response no longer includes the relevant next-page token.
• Keep v on every continuation request. The continuation token selects the underlying video content.
• Tokens containing percent sequences can be passed as returned or URL-encoded with curl -G --data-urlencode.

Error Handling

Notes

• Private, removed, age-gated, or region-blocked videos can return 502 failed to parse video results and are not charged.
• Optional fields are omitted when YouTube does not provide them; one unavailable field does not fail an otherwise valid response.
• Display strings follow hl, while extracted_* fields are normalized to integers when the value can be determined safely.
• Match hl and gl, such as hl=fr&gl=fr, for consistent localized metadata and regional ranking.