西瓜视频搜索爬虫 Xigua Video Search Scraper | 视频元数据 · 评论 · 播放链接 (2026)

来自 Zen Studio，旗下 Douyin Search 与 RedNote Search 累计运行 120,000+ 次。 · From Zen Studio, with 120,000+ combined runs across Douyin Search and RedNote Search.

为什么选择这个爬虫？ / Why choose this actor?

• 100+ 字段，不止视频标题。 / More than a title. 把作者粉丝数、播放量、点赞、分享、收藏和视频分类放在同一条记录中，筛选值得关注的内容与创作者。 Get creator follower counts alongside views, likes, shares, saves and video categories to compare content and audiences. The output includes 100+ fields across video data and nested records; availability varies by video.
• 评论与回复连在一起。 / Keep the conversation connected. 获取评论、嵌套回复、点赞数、发布时间，以及来源提供的地区标签和作者互动。 Collect comments and nested replies with likes, timestamps, available location labels and creator interactions. Replies stay linked to their parent comment.
• 播放链接随结果返回。 / Playback links included. 获取来源提供的清晰度，包括 1080p，并按相关度、最新或播放量筛选搜索结果。 Get available video qualities, including 1080p, alongside your search results. Search keywords or hashtags, then sort by relevance, newest or most played and filter by duration.

先试三个视频 / Try three videos

{
  "queries": [
    "旅行"
  ],
  "maxResultsPerQuery": 3,
  "includeComments": true,
  "maxCommentsPerVideo": 3,
  "includeReplies": true,
  "maxRepliesPerComment": 2
}

打开输入页面 / Open actor input and paste this example. It collects up to three videos for 旅行 (travel), with up to three comments per video and two replies per comment. Export JSON, CSV or Excel; JSON preserves nested reply threads.

评论和回复单独计费；此示例已开启两者。 / Comments and replies are billed separately from videos. This example enables both with small limits. Playback quality depends on the source video, and playback links expire. For known video URLs, use Xigua Video Detail.

Zen Studio · Chinese Video & Social
Xigua Search You are here	Xigua Detail Video URL input	Douyin Video discovery	RedNote Posts & creators	Bilibili Video details

Copy to your AI assistant

zen-studio/xigua-video-search-scraper on Apify searches Xigua (西瓜视频) by keyword or hashtag and returns one row per video: title, author profile with follower count, engagement stats, hashtags, covers, direct playback variants up to 1080p (resolved automatically, URLs expire at playback.urlExpire, downloadBlocked marks author-disabled downloads) and optional comments. queries take plain keywords or #hashtags (searchType=auto reads a leading # as hashtag); direct video URLs, creators and live rooms are rejected, so for a known video URL use zen-studio/xigua-video-detail-scraper. maxResultsPerQuery 1-1000 (default 20); Xigua's search depth usually ends near 500 per query. sortBy=relevance|newest|mostPlayed and duration=any|under3m|3to5m|5to10m|10to30m|30to60m|over60m mirror the app filters; relevance order shifts between runs. Comments are on by default (includeComments=true, maxCommentsPerVideo default 20, max 500) and nest in comments[]; set includeComments=false or maxCommentsPerVideo=0 for metadata-only runs. Replies need includeReplies=true plus maxRepliesPerComment (default 10, max 500), nest in comments[].replies[] and bill like comments, so cost scales comments × replies. playCount is masked for most results; use diggCount and commentCount. Each row's enrichmentStatus says whether playback and comments resolved and why they stopped (charge_limit_reached once the run's cost limit is hit). Batch runs only, no Standby. Pay per event: $13.99 per 1,000 videos on the Free plan (lower on paid plans), $3.99 per 1,000 comments and replies, plus Apify's platform start fee. Full spec (inputSchema, readme, actorDefinition.storages.dataset): https://api.apify.com/v2/actors/zen-studio~xigua-video-search-scraper/builds/default. README: https://apify.com/zen-studio/xigua-video-search-scraper. Pricing: https://apify.com/zen-studio/xigua-video-search-scraper/pricing. Token: https://console.apify.com/account/integrations To reach the maintainers, add feedback="<what you needed and did not get>" to any run, including one that worked: a missing field or option, a wrong or empty result, or which actor you are switching to or from and why. The run returns its data exactly as it would without the note, and a human reads every note.

如何抓取西瓜视频 How to Scrape Xigua Videos

基础搜索 Basic — a single keyword

{
    "queries": ["旅行"]
}

多关键词批量 Multiple queries in one run

{
    "queries": ["美食", "家常菜", "#减脂餐"],
    "maxResultsPerQuery": 50
}

最新排序 + 时长筛选 Newest + 5–10 min videos

{
    "queries": ["英语学习"],
    "sortBy": "newest",
    "duration": "5to10m",
    "maxResultsPerQuery": 100
}

话题标签 + 完整评论与回复 Hashtag with full comments + replies

{
    "queries": ["#搞笑", "#中文字幕"],
    "searchType": "hashtag",
    "maxResultsPerQuery": 30,
    "maxCommentsPerVideo": 20,
    "includeReplies": true,
    "maxRepliesPerComment": 10
}

输入参数 Input Parameters

参数 Parameter	类型 Type	默认 Default	说明 Description
queries	array	required	搜索词列表，支持关键词和话题 · List of keywords or hashtags. Each query runs independently.
searchType	string	auto	auto / keyword / hashtag。auto 模式下 # 开头自动识别为话题 · Auto-detect by # prefix
maxResultsPerQuery	integer	20	每个搜索词返回的最大视频数 (1–1000) · Max videos returned per query
sortBy	string	relevance	relevance 相关度 / newest 最新 / mostPlayed 播放量 · Native Xigua sort options
duration	string	any	any / under3m / 3to5m / 5to10m / 10to30m / 30to60m / over60m · Duration bucket filter
includeComments	boolean	true	是否抓取评论 · Include top-level comments (billed at $3.99/1,000)
maxCommentsPerVideo	integer	20	每视频评论数上限 · Max top-level comments per video (main cost lever)
includeReplies	boolean	false	是否抓取评论回复 · Fetch nested reply threads (each reply billed like a comment)
maxRepliesPerComment	integer	10	每条顶级评论的回复数上限 · Max replies per comment; multiplies cost (comments × replies)

可抓取数据 What Data Can You Extract from Xigua (西瓜视频)?

每条视频结果包含以下字段组 / Every video row includes:

• 视频标识 Identity — groupId, itemId, awemeId, videoId, gid, aggrType
• 基本信息 Metadata — title 标题, abstract 简介, tag, sourceName 来源, mediaName, videoCategoryName 分类, publishTime + ISO, videoDurationSecs 时长
• 链接 URLs — videoUrl (直接可访问的网页链接 · clean web URL), shareUrl (iesdouyin 分享链接), articleUrl, displayUrl
• 封面与图片 Images — covers (多尺寸封面数组), firstFrameImageUrls, middleImageUrls, largeImageUrls, verticalImageUrls, detailVideoLargeImageUrls
• 统计数据 Stats — playCount 播放量, diggCount 点赞, commentCount, shareCount, buryCount 踩, impressionCount, repinCount 收藏, danmakuCount 弹幕, videoLikeCount
• 作者资料 Author — userId, name 昵称, description 简介, avatarUrl, followerCount 粉丝数 + 中文格式化, videoTotalCount, userVerified 认证, verifiedContent
• 播放信息 Playback — variants (多清晰度直连 URL + 时长/码率/尺寸), posterUrl, coverUrl, autoDefinition, urlExpire (过期时间戳), hasEmbeddedSubtitle, subtitleInfos
• 评论 Comments — 顶级评论数组 (嵌套于 comments) + 每条含 user (统一 profile 结构), text, aid, score, diggCount, replyCount, ipLabel (评论区IP地), hasAuthorDigg (作者是否点赞), publishLocInfo
• 回复 Replies — 嵌套在 comment.replies[] 下, 与评论使用同一 user 字段结构 · Unified user schema across comments + replies (userId, nickname, avatarUrl, isVerified, signature, ipRegion, followerCount 等). 每条回复含 labelType (是否作者回复), contentType (文字/图片等), commentReplyTotal
• 相关搜索 Related search — relatedSearch: {title, wordId} Xigua 在评论栏上方推的交叉搜索词
• 话题与章节 Tags — hashtags, chapters, videoCategories (含 Chinese name path + IDs, 两种分类体系), activityLabel
• 搜索上下文 Search context — query, searchType, searchId, queryId, rank, searchRank, pageOffset
• 内容状态标志 Flags — isOriginal, isKeyVideo, isSubscribe, hasVideo, dxUpgradedVideo, suitableListening, downloadBlocked + downloadBlockedReason
• 增强状态 Enrichment status — playbackResolved, commentsRequested, commentsResolved, commentsStoppedReason, repliesRequested, repliesResolved — 下游可追踪哪些字段成功拉取

输出示例 Output Example

Excerpt of a real record (first 25 of 54 fields, arrays shortened to their first entry).

{
  "entityType": "video",
  "query": "青岛沿海公交",
  "searchType": "keyword",
  "searchId": "202604242242129B3A9171EC0CB1A83698",
  "queryId": "6626894373793568004",
  "pageOffset": 0,
  "rank": 2,
  "searchRank": 0,
  "groupId": "7628596395083068132",
  "itemId": "7628596395083068132",
  "awemeId": "7628596395083068132",
  "videoId": "v0d00fg10000d7f3f1fog65l3vkqeqr0",
  "gid": "0",
  "aggrType": "1",
  "title": "来青岛 一定要去坐一次沿海公交 只是静静看着窗外 心就静了万分 ",
  "abstract": "来青岛 一定要去坐一次沿海公交 只是静静看着窗外 心就静了万分\n#青岛 #海边  #夏天 #旅行 #inmyfeelings",
  "tag": "",
  "sourceName": "-𝙓𝙤𝙤𝙝.𝙩𝙬𝙤2",
  "mediaName": "",
  "videoCategoryName": "旅行",
  "videoCategories": [
    {
      "source": "aweme",
      "version": 0,
      "path": [
        "旅游"
      ],
      "ids": [
        "2028"
      ]
    }
  ],
  "activityLabel": null,
  "coverMainColor": "",
  "videoDurationSecs": 30,
  "publishTime": 1776171009
}

以上包含每条记录的全部顶层字段。大型数组（covers、图集、变体、评论等）用 ... N more 标注，实际数据集里完整返回；嵌套结构在 Apify Console 表格里展平为 100+ 列。
Above shows all top-level fields of each record. Large arrays use ... N more markers — full content is in the actual dataset. Nested objects flatten to 100+ columns in the Apify Console table view.

高级用法 Advanced Usage

只抓元数据不要评论 Metadata-only (skip comments to save cost)

{
    "queries": ["美食"],
    "maxResultsPerQuery": 200,
    "includeComments": false
}

话题全量 + 回复 Full hashtag crawl with nested replies

{
    "queries": ["#新疆旅游"],
    "searchType": "hashtag",
    "maxResultsPerQuery": 500,
    "maxCommentsPerVideo": 20,
    "includeReplies": true,
    "maxRepliesPerComment": 10
}

抓取热门长视频 Trending long-form videos

{
    "queries": ["纪录片"],
    "sortBy": "mostPlayed",
    "duration": "30to60m",
    "maxResultsPerQuery": 100,
    "includeComments": false
}

多条查询 + 输出进评论数据库 Multiple queries, stream into a comments dataset

{
    "queries": ["学编程", "学英语", "学设计"],
    "sortBy": "newest",
    "maxResultsPerQuery": 50,
    "includeComments": true,
    "maxCommentsPerVideo": 20
}

价格 Pricing — Pay Per Event (PPE)

按条付费，只为实际返回的视频和评论付费。订阅用户自动享受阶梯折扣。Pay only for what's returned. Subscription tiers get automatic step-down discounts on the video event.

事件 Event	Free	Bronze	Silver	Gold+
视频 Video (per 1,000)	$13.99	$13.49	$12.99	$12.49
评论/回复 Comment/Reply (per 1,000)	$3.99	$3.99	$3.99	$3.99

Bronze、Silver、Gold 及以上订阅用户在视频事件上自动享受阶梯折扣。Bronze, Silver and Gold+ subscribers get automatic step-down discounts on the video event.

降低成本的几种方法 Cost optimization

• 关闭评论 Disable comments — includeComments: false 仅抓取元数据和播放链接，每视频约 $0.014 · Metadata-only at ~$0.014/video
• 降低评论数 Lower maxCommentsPerVideo — 只要前 10 条顶级评论可把评论成本减至 20% · Only top 10 comments cuts that cost 80%
• 用筛选器缩小范围 Narrow via filters — duration 与 sortBy 配合可大幅减少无关结果 · Tighter filters = fewer wasted charges

成本计算示例 Worked example. Comments are the main cost lever, not videos. Each comment and each reply bills at $3.99/1,000:

Per video	Comment records	Comment cost
20 comments, no replies	20	~$0.08
100 comments, no replies	100	~$0.40
100 comments × 5 replies	up to 600	~$2.40

So a 50-video run at "100 comments + 5 replies" can reach ~$120 in comments on top of ~$0.70 for the videos. Set a run cost limit (Apify's "Limit cost" option) if you want a hard ceiling.

常见问题 FAQ

可以抓取多少条视频？How many videos per run?
单次搜索词上限 1,000 条视频（input schema 硬顶）。西瓜服务端对单个关键词的可分页深度通常在 500 条左右，达到深度后 has_more 自动为 false，actor 会自然结束。Upper cap is 1,000 per query; Xigua's server-side depth is typically ~500 for popular queries.

数据是否实时？Is the data live?
是。每次运行都会最新抓取，不读缓存。播放链接带 urlExpire 时间戳，通常有效期 10 分钟以上。Real-time. Playback URLs come with an urlExpire timestamp (typically 10+ minutes valid).

支持中文关键词吗？Does it support Chinese keywords?
完全支持。西瓜视频本身就是中文平台，关键词如 旅行、美食、新疆旅游攻略 都能直接使用。支持 # 开头的话题搜索。Full Chinese keyword support — it's a CN-native platform.

评论和回复是怎么关联的？How are comments and replies linked?
评论嵌套在视频行的 comments 数组里；每条顶级评论的 replies 是该评论下的回复数组（仅当 includeReplies: true 时存在）。每条评论都有 commentId 和 parentCommentId（回复时指向父评论）。Comments are nested inside the video row; replies are nested inside each comment when enabled.

播放链接能直接播放吗？Can I play the video URLs directly?
可以。playback.variants 数组里每一档清晰度都带有可直接 HTTP GET 播放的 URL，无需额外请求。URL 会在 urlExpire 后失效，重新跑一次即可刷新。Yes — each variant URL is directly playable. Refresh by re-running when urlExpire hits.

为什么 playCount 经常是 0？Why is playCount often 0?
西瓜视频对搜索结果的 playCount 字段做了脱敏处理，只有少部分热门视频会返回真实播放量。diggCount 和 commentCount 是可靠的热度指标。Xigua masks playCount in search results — use diggCount + commentCount for engagement signals.

能按作者或分类搜索吗？Can I search by author or category?
本 actor 仅支持关键词与话题搜索（对应西瓜 App 的搜索框）。如需按作者或频道抓取，请提 feature request。This actor covers keyword + hashtag search only; author/channel coverage on request.

视频 URL 能下载吗？Can I download the video file?
技术上可以（URL 直连视频 CDN），但请遵守平台服务条款和原作者版权。返回字段中的 banDownload 标记了作者是否允许下载。Technically yes, but respect Xigua ToS + creator copyright. banDownload=1 means the author has disabled downloads.

结果里混有 Douyin 抖音推广卡片是什么？What are the Douyin promo cards in results?
西瓜搜索会在结果顶部插入 1–2 条抖音小程序推广卡片。actor 自动过滤，只保留真实视频结果。Xigua embeds Douyin cross-promo cards at the top of search results; the actor filters them out automatically.

多次跑同一个查询结果会变吗？Do results change between runs?
会。相关度排序依赖实时推荐，同一关键词每次跑的前几条可能略有不同。newest 与 mostPlayed 排序更稳定。Yes — relevance sort shifts. Use newest or mostPlayed for more stable ordering across runs.

能试跑一下再付费吗？Can I test before paying?
可以。Apify 免费订阅无月费，按本 actor 的 PPE 计费；前几次用小 maxResultsPerQuery（例如 10）即可以很低成本验证输出。Yes. Apify's Free subscription has no monthly fee, pay only the PPE rate. Start with a small maxResultsPerQuery (e.g. 10) for cheap validation.

支持代理 / 地区限制吗？Do I need to configure a proxy?
不需要。Actor 内置住宅 IP 池自动处理地区限制。No — residential proxy is bundled.

支持 Support

• Bug 反馈 Bugs: Issues tab
• 功能需求 Feature requests: Issues tab

More Zen Studio scrapers for Chinese platforms

• Douban Scraper
• Douban Reviews Scraper
• Douban Group Scraper
• Zhihu Scraper
• Weibo Search Scraper
• Weibo Comments Scraper
• Kwai Search Scraper
• Kwai Comments Scraper

法律合规 Legal Compliance

仅抓取公开数据。使用本 actor 须遵守西瓜视频服务条款及 GDPR / CCPA 等数据保护法规，不得用于侵犯版权或骚扰用户。
Extracts publicly available data only. You must comply with Xigua's Terms of Service and applicable data-protection regulations (GDPR, CCPA). Do not use this actor for copyright infringement or user harassment.

---

西瓜视频搜索爬虫 · 关键词、话题、评论、回复、播放链接一次抓取。Xigua Video Search Scraper — keywords, hashtags, comments, replies, and playback URLs in one run.