[中文]

用途:

• 返回某条一级评论下的二级回复。
• 链路：先调 fetch_article_comments 取到每条评论的 content_id，再把它喂给本接口。
• 价格：0.01$/次
• ⏱️ 由于微信服务器原因，本接口响应较慢，请将客户端请求超时（timeout）设置为 30 秒；timeout 设置过小会造成已扣费但收不到响应的情况。
• ⚠️ 大整数 ID 精度：响应中的 content_id / reply_id 等为 64 位大整数，超出 JavaScript 安全整数范围（2^53-1）。请始终以字符串方式接收 / 传递这类 ID（解析 JSON 用 json-bigint 或按文本取值），勿让其经过 JS Number。Swagger UI 文档页对超大整数会末位舍入显示，属正常现象，不影响接口实际返回的数据。
• 💡 提速：传 comment_id 可让上游跳过整篇文章的抓取解析，实测本接口从 15.10s 降到 6.91s / 评论接口从 8.28s 降到 6.13s。该值取自 /fetch_article_detail 或 /fetch_article_detail_h5 响应的 $.data.content.comment_id，拿它是零成本的。

参数:

• url: 公众号文章链接（https://mp.weixin.qq.com/s/…）。示例 http://mp.weixin.qq.com/s?__biz=Mzk3NTA0MzM5NA==&mid=2247483745&idx=1&sn=3f34e768cf457a501038991ed30be1f4#rd
• content_id: 可选（纯数字）。目标一级评论的 content_id，取自 fetch_article_comments 响应的 $.data.comments[N].content_id；不传则自动取第一条有回复的评论。示例 12109128638545265979
• offset: 可选，默认 0（>=0）。回复翻页偏移；首页 0，has_more=true 时传上次返回的 next_offset 取下一页。
• ⚠️ 上游已移除 all_pages 自动翻页：它一次计费最多打 30 个上游请求，且自动翻页模式不返回 has_more / next_offset，调用方无从发现结果被截断（实测 400 条回复只给回前 300）。翻页统一用 offset / next_offset 驱动，与 /fetch_article_comments 的游标契约一致。
• raw: 可选，默认 True。True=完整原始回复对象；False=精简回复。

返回:

• 二级回复列表及翻页信息

注意：本接口 raw=true 与 raw=false 的外层结构一致，差异仅在 replies[] 每个元素的字段粒度（精简视图 vs 完整原始对象）。

响应结构与 JSON Path:

raw=false（精简回复）:

• $.data.content_id: 目标评论 id（回显）
• $.data.parent_nick: 父评论（被回复者）昵称
• $.data.reply_total: 二级回复总数
• $.data.count: 本次返回的回复条数
• $.data.replies[N].reply_id: 回复 id
• $.data.replies[N].nick_name: 回复者昵称
• $.data.replies[N].content: 回复内容
• $.data.replies[N].create_time: 回复时间戳
• $.data.replies[N].ip_wording: IP 归属地
• $.data.replies[N].identity_name: 身份标识
• $.data.replies[N].from_ai / reply_is_elected: 是否 AI 回复 / 是否精选
• 翻页字段: $.data.offset / $.data.next_offset / $.data.has_more

raw=true（原始）:

• 外层 content_id / parent_nick / reply_total / count（及单页模式的 offset / next_offset / has_more）与 raw=false 完全一致；差异仅在 replies[] 元素为完整原始回复对象（含 raw=false 投影掉的字段）。常用路径仍为 $.data.replies[N].content / $.data.replies[N].reply_id / $.data.replies[N].nick_name。

[English]

Purpose:

• Get the second-level replies under a first-level comment.
• Chain: first call fetch_article_comments to get each comment's content_id, then feed it into this endpoint.
• Price: $0.01 per request
• ⏱️ Due to WeChat server latency, this endpoint responds slowly; please set your client request timeout to 30 seconds — a timeout that is too small may result in being billed without receiving the response.
• ⚠️ Large-integer ID precision: IDs such as content_id / reply_id in the response are 64-bit big integers beyond JavaScript's safe-integer range (2^53-1). Always receive / pass such IDs as strings (parse JSON with json-bigint or read them as text), never through JS Number. Swagger UI rounds the trailing digits of huge integers in its docs view — this is expected and does not affect the actual data returned by the API.

Parameters:

• url: WeChat MP article URL (https://mp.weixin.qq.com/s/…). E.g. http://mp.weixin.qq.com/s?__biz=Mzk3NTA0MzM5NA==&mid=2247483745&idx=1&sn=3f34e768cf457a501038991ed30be1f4#rd
• content_id: Optional (numeric). content_id of the target first-level comment, from $.data.comments[N].content_id of the fetch_article_comments response; defaults to the first comment that has replies when omitted. E.g. 12109128638545265979
• offset: Optional, default 0 (>=0). Reply pagination offset; 0 for the first page, pass next_offset from the previous response when has_more=true.
• Note: all_pages was removed upstream — it issued up to 30 upstream requests per billed call and omitted has_more / next_offset, so callers could not tell the result was truncated. Paginate with offset / next_offset instead.
• raw: Optional, default True. True=full raw reply objects; False=simplified replies.

Return:

• Second-level reply list with pagination info

Note: the outer structure is identical for raw=true and raw=false; the difference is only in the field granularity of each replies[] element (simplified view vs full raw object).

Response structure & JSON Path:

raw=false (simplified replies):

• $.data.content_id: target comment id (echo)
• $.data.parent_nick: nickname of the parent comment (the one replied to)
• $.data.reply_total: total second-level replies
• $.data.count: number of replies returned this time
• $.data.replies[N].reply_id: reply id
• $.data.replies[N].nick_name: replier nickname
• $.data.replies[N].content: reply content
• $.data.replies[N].create_time: reply timestamp
• $.data.replies[N].ip_wording: IP region
• $.data.replies[N].identity_name: identity label
• $.data.replies[N].from_ai / reply_is_elected: AI reply / featured flags
• Pagination fields: $.data.offset / $.data.next_offset / $.data.has_more

raw=true (raw):

• Outer content_id / parent_nick / reply_total / count (and offset / next_offset / has_more in single-page mode) are identical to raw=false; the only difference is that replies[] elements are full raw reply objects (including fields projected away in raw=false). Common paths remain $.data.replies[N].content / $.data.replies[N].reply_id / $.data.replies[N].nick_name.

{
  "url": "http://mp.weixin.qq.com/s?__biz=Mzk3NTA0MzM5NA==&mid=2247483745&idx=1&sn=3f34e768cf457a501038991ed30be1f4#rd",
  "content_id": "12109128638545265979"
}