先决条件
- 一个 Bright Data 账户，具有有效的 API 密钥
- 完成了 快速入门
请求结构
每个同步请求都遵循相同的模式：POST https://api.brightdata.com/datasets/v3/scrape?dataset_id={DATASET_ID}&format=json
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json
[{"url": "https://www.youtube.com/..."}]
dataset_id 和输入 URL 格式。
/scrape 端点专为实时的单 URL 请求设计。Bright Data 建议每个请求不超过 20 个 URL（这是最佳实践，而非硬性限制），端点最多等待 1 分钟。如果采集耗时更长，响应为 HTTP 202，返回 snapshot_id 而不是数据，采集将继续以异步方式进行。请通过异步请求获取结果，或使用 Webhook 投递接收结果。频道
数据集 ID：gd_lk538t2k2p1k3oos71
curl -X POST \
"https://api.brightdata.com/datasets/v3/scrape?dataset_id=gd_lk538t2k2p1k3oos71&format=json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '[{"url": "https://www.youtube.com/@MrBeast"}]'
200 响应。处理时间取决于目标网站状况和爬虫负载。
示例响应
示例响应
[
{
"subscribers": 516000000,
"url": "https://www.youtube.com/@mrbeast",
"handle": "@MrBeast",
"handle_md5": "dd748d5dfc88b91ea388429c3c685874",
"banner_img": "https://yt3.googleusercontent.com/mHMO_eEMp0dPvh0ADwXhPXNYb_GnjSVsLI8biqF1CpxT8OPl7izhNQsDPD3JHhd5y5Mg9GrP=w2560-fcrop64=1,00005a57ffffa5a8-k-c0xffffffff-no-...",
"profile_image": "https://yt3.googleusercontent.com/nxYrc_1_2f77DoBadyxMTmv7ZpRZapHR5jbuYe7PlPd5cIRJxtNNEYyOC0ZsxaDyJJzXrnJiuDE=s160-c-k-c0x00ffffff-no-rj",
"name": "MrBeast",
"Description": "SUBSCRIBE FOR A COOKIE!\nNew MrBeast or MrBeast Gaming video every single Saturday at noon eastern time!\nAccomplishments:\n- Raised $20,000,000 To Plant 20,000...",
"videos_count": 1000,
"created_date": "2012-02-19T00:00:00.000Z"
}
]
YouTube 频道抓取器返回哪些输出字段？
YouTube 频道抓取器（gd_lk538t2k2p1k3oos71，collect_by_url）返回 21 个输出字段。字段说明保留 API 返回的英文原文。
输出字段（21）
输出字段（21）
| 字段 | 类型 | 说明 |
|---|---|---|
url | url | URL that link directly to the YouTube profile |
handle | text | Unique to each user (e.g., @name) |
handle_md5 | text | This field stores the MD5 hash of the handle field. |
banner_img | url | URL that links to the banner photo |
profile_image | url | URL that link to the profile picture |
name | text | Profile name |
subscribers | number | How many subscribers the profile have |
Description | text | Brief descriptions or statements that users write to introduce themselves or share their interests and passions on YouTube |
videos_count | number | How many videos the profile have |
created_date | date | The date the profile was created |
views | number | How many views the profile have |
Details | object | channel Details |
Links | array | External websites that users may link to from their YouTube profiles |
identifier | text | identifier |
discovery_input | object | Discovery input value |
id | text | channel id |
has_podcast | boolean | has podcast indicator |
top_videos | array | top videos |
featured_video | url | Featured Video |
featured_channels | array | featured channels |
collaborations | array | Collaborations videos |
/scrapers/gd_lk538t2k2p1k3oos71.json
视频
数据集 ID：gd_lk56epmy2i5g7lzu0k
curl -X POST \
"https://api.brightdata.com/datasets/v3/scrape?dataset_id=gd_lk56epmy2i5g7lzu0k&format=json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '[{"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"}]'
示例响应
示例响应
[
{
"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
"title": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
"views": 1812511291,
"likes": 19374379,
"date_posted": "2009-10-25T06:57:33.000Z",
"description": "The official video for “Never Gonna Give You Up” by Rick Astley. \n\nNever: The Autobiography 📚 OUT NOW! \nFollow this link to get your copy and listen to Rick...",
"num_comments": 2457228,
"youtuber": "@RickAstleyYT",
"youtuber_md5": "4fc65cd6213b5c0f1d04c2aa3006f46a",
"video_url": "https://rr2---sn-n4v7snse.googlevideo.com/videoplayback?expire=1788712345&ei=OUGdaoSULZ-RsfIP8OmfsAo&ip=208.41.164.3&id=o-ALklNMQDHiVh32KdRiuWf9qyPo-8mNxywN-..."
}
]
YouTube 视频抓取器返回哪些输出字段？
YouTube 视频抓取器（gd_lk56epmy2i5g7lzu0k，collect_by_url）返回 49 个输出字段。字段说明保留 API 返回的英文原文。
输出字段（49）
输出字段（49）
| 字段 | 类型 | 说明 |
|---|---|---|
url | url | The full YouTube watch URL of the collected video |
title | text | Title of the YouTube video |
youtuber | text | The channel handle extracted from the channel’s browse endpoint |
youtuber_md5 | text | MD5 hash of the youtuber field |
video_url | url | Direct streaming URL of the video file, extracted from YouTube’s adaptive formats or decoded from signatureCipher |
video_length | number | Duration of the video in seconds |
likes | number | Number of likes received by the video |
views | number | Number of views the video has received |
date_posted | date | Date when the video was published |
description | text | Text description accompanying the video |
num_comments | number | Number of comments on the video, as displayed by YouTube (includes both top-level comments and replies in YouTube’s count) |
subscribers | number | Number of subscribers to the YouTube channel |
music | object | Information about the music used in the video. Only present when YouTube’s Content ID system has identified music in the video |
video_id | text | Unique identifier for the YouTube video |
channel_url | url | URL-encoded URL of the YouTube channel, derived from the channel’s canonical browse endpoint |
preview_image | image | URL of the preview image for the video |
shortcode | text | The unique short identifier of the video, identical to video_id, used across all YouTube video types including regular videos and Shorts |
verified | boolean | Indicates whether the channel is verified |
handle_name | text | The display name of the YouTube channel as shown on the video page |
avatar_img_channel | url | The URL of the avatar image of the YouTube channel |
is_sponsored | boolean | Indicates whether the video contains a paid content overlay as declared by YouTube |
related_videos | array | List of URLs of videos shown in the sidebar alongside the current video |
license | url | URL of the video’s license as declared in YouTube’s video metadata |
viewport_frames | text | Resolution of the primary video stream format, derived from YouTube’s streaming data |
current_optimal_res | text | Current and optimal video resolution with frame rate, from YouTube’s streaming format metadata |
codecs | text | Video and audio codec strings formatted as video codec and audio codec, extracted from YouTube’s adaptive streaming formats |
color | text | Color primaries of the video stream, extracted from YouTube’s adaptive format color info |
quality | text | Quality identifier of the first adaptive streaming format as reported by YouTube |
quality_label | text | Human-readable video quality label as reported by YouTube |
post_type | text | Type of YouTube content inferred from the URL structure. Possible values: post for standard video, short for YouTube Short, or N/A |
youtuber_id | text | YouTube’s internal channel ID, distinct from the channel handle |
transcript | text | Full plain-text transcript of the video, compiled from YouTube’s caption tracks. English captions are preferred, falls back to the first available language if English is unavailable |
formatted_transcript | array | Timestamped transcript segments from YouTube’s caption tracks, each containing start time, end time, and duration in milliseconds, and the caption text |
hashtags | array | List of hashtags found in the video title or description, each with the hashtag text and its YouTube search link |
tags | array | Tags of the video |
next_recommended_videos | array | Videos shown as end-screen cards during the final seconds of the video, including timing in milliseconds and display style |
recommended_videos | array | Videos recommended by YouTube in the sidebar alongside the current video |
transcript_language | array | List of caption and transcript languages available for this video on YouTube, including whether each track was auto-generated |
chapters | array | List of chapter markers defined by the creator, each with a title, start time in milliseconds, and preview image |
transcription_language | text | The language code requested by the caller for transcript retrieval, passed as an input parameter at collection time |
is_age_restricted | boolean | Whether the video is age-restricted, inferred from a non-OK playability status combined with an age-related reason from YouTube |
channel_url_decoded | text | Plain-text percent-decoded version of the channel URL |
category | text | YouTube’s content category assigned to the video, from YouTube’s predefined category taxonomy |
audio_tracks | array | audio Tracks |
made_for_kids | boolean | Indicates whether the video is designated as made for kids, inferred from the YouTube miniplayer playback mode |
player_notification_text | text | User-facing notification text shown inside the video player |
ai_content_label | boolean | Indicates whether the video has been labeled as AI-generated or significantly AI-altered, either by the creator or auto-detected by YouTube. |
live_status | text | Livestream status of the video, equivalent to yt-dlp’s live_status. One of “is_live”, “was_live”, or “not_live”, parsed from YouTube’s info string (e.g. “Streamed live 29 minutes ago”) or defaulted to “not_live” for regular videos |
comments | array | 10 Most relevant comments of the video |
/scrapers/gd_lk56epmy2i5g7lzu0k.json
评论
数据集 ID：gd_lk9q0ew71spt1mxywf
curl -X POST \
"https://api.brightdata.com/datasets/v3/scrape?dataset_id=gd_lk9q0ew71spt1mxywf&format=json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '[{"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"}]'
示例响应
示例响应
[
{
"url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
"likes": 311000,
"replies": 0,
"comment_id": "Ugzge340dBgB75hWBm54AaABAg",
"comment_text": "can confirm: he never gave us up",
"username": "@Yo***e",
"username_md5": "0e4d7a0778633b4acf6631dd251f0797",
"user_channel": "https://www.youtube.com/@YouTube",
"date": "1 year ago",
"video_id": "dQw4w9WgXcQ"
}
]
YouTube 评论抓取器返回哪些输出字段？
YouTube 评论抓取器（gd_lk9q0ew71spt1mxywf，collect_by_url）返回 16 个输出字段。字段说明保留 API 返回的英文原文。
| 字段 | 类型 | 说明 |
|---|---|---|
comment_id | text | Unique identifier for each comment |
comment_text | text | The text content of the comment |
likes | number | Number of likes received by the comment |
replies | number | Number of replies to the comment |
username | text | Username of the commenter |
username_md5 | text | This field stores the MD5 hash of the username field. |
user_channel | text | URL of the commenter’s YouTube channel |
date | text | Date when the comment was posted |
url | url | Web address of the YouTube video where the comment was posted |
video_id | text | Unique identifier for the YouTube video |
replies_value | array | Array of reply objects associated with this comment |
replies_without_names | array | Array of reply objects with personally identifiable information (usernames) removed, containing only MD5 hashed user identifiers |
user_id | text | Unique identifier for the commenter’s YouTube channel |
mentioned_timestamps_minutes | array | Array of timestamps mentioned in the comment, formatted in minutes (e.g., ‘2:35’) |
mentioned_timestamps_seconds | array | Array of timestamps mentioned in the comment, converted to total seconds |
date_iso | date | Date in ISO format |
/scrapers/gd_lk9q0ew71spt1mxywf.json
快速参考：数据集 ID
| 端点 | 数据集 ID | URL 模式 |
|---|---|---|
| 视频 | gd_lk56epmy2i5g7lzu0k | youtube.com/watch?v={video_id} |
| 频道 | gd_lk538t2k2p1k3oos71 | youtube.com/@{handle} |
| 评论 | gd_lk9q0ew71spt1mxywf | youtube.com/watch?v={video_id} |
输出格式
使用format 查询参数控制响应格式：
| 值 | 描述 |
|---|---|
json | JSON 数组 |
ndjson | 换行分隔的 JSON，每行一条记录（默认） |
csv | 逗号分隔值 |
/scrape 的 format 值。完整列表（NDJSON（默认）、JSON、JSON lines、CSV 和 Parquet）及各格式适用范围见输出格式。
相关内容
异步批量请求
在单个批量作业中抓取数百个 URL。
API 参考
完整的参数和响应字段参考。