# 如何批量获取一条推文的回复 用 2 个接口：X 帖子详情 → X 帖子回复列表。下面的示例一共 4 次调用，合计 $0.0063；失败的调用不收费。 拿到一条 X（推特）帖子下面的回复，按 X 自己的相关度一页一页往下翻：正文、作者、发布时间和互动数，再加上这条帖子本身的数据。 ## 一步一步做 ### 第 1 步：X 帖子详情 按帖子编号，或按链接（传 `url`，x.com 或 twitter.com 上带 /status/ 的链接）取这条帖子。`metrics.comments` 是回复数，下一步用它的 `id`。（[X 帖子详情](https://dataxapi.com/zh/apis/x/posts/get/)，$0.0003 / 一次） cURL ```bash curl -G "https://api.dataxapi.com/v1/x/posts/get" \ --data-urlencode "id=2108198226532708497" \ -H "Authorization: Bearer $DATAXAPI_API_KEY" ``` Python ```python from dataxapi import DataXAPI dx = DataXAPI() # reads DATAXAPI_API_KEY result = dx.x.posts.get(id="2108198226532708497") print(result["data"]["item"]) ``` TypeScript ```typescript import { DataXAPI } from "@dataxapi/sdk"; const dx = new DataXAPI(); // reads DATAXAPI_API_KEY const result = await dx.x.posts.get({ id: "2108198226532708497" }); console.log(result.data.item); ``` 命令行 ```bash dataxapi call x.posts.get id=2108198226532708497 ``` 录制的返回：取自 2026-10-12 19:04 UTC 录下的真实返回。名字、正文和编号是示例，链接也是，都不是录到的原文（页面不展示任何人的个人信息）；数字是原样。长列表做了截断。请求号、游标和余额是示意。 ```jsonc { "code": "ok", "request_id": "9a3158fb-2f1c-4829-9fcc-9b9690d84bb5", "data": { "item": { "id": "1889020583051160744", "url": "https://x.com/i/status/1889020583051160744", "type": "video", "text": "Three things nobody tells you about weekend hiking.", "lang": "en", "published_at": "2026-10-08T14:10:00.000Z", "author": { "id": "171254161496063062", "username": "maya_makes", "name": "Northside Coffee", "avatar_url": "https://pbs.twimg.com/profile_images/1817170446354046662/qJUxX-tg_normal.jpg", "verified": true, "followers": 932620 }, "metrics": { "likes": 155, "comments": 7, "favorites": 11, "views": 10638, "reposts": 40 }, "images": [], "hashtags": [], "mentions": [ { "id": "192596166327376049", "username": "priya_reads", "name": "Priya R." } ], "post_id": null, "parent_id": null } }, "warnings": [], "meta": { "capability": "x.posts.get", "result_status": "complete", "fetched_at": "2026-10-12T19:04:36.000Z", "cached": false, "replayed": false, "unavailable": { "title": "platform", "metrics.shares": "platform", "is_pinned": "capability", "root_id": "platform" }, "billing": { "amount": "0.0003", "currency": "USD", "balance_after": "9.9997" } } } ``` ### 第 2 步：X 帖子回复列表 按帖子 `id` 取回复。X 的回复本身就是帖子，按帖子返回。翻页时把 `page.next_cursor` 作为 `cursor` 传进去，最后一页不带游标。（[X 帖子回复列表](https://dataxapi.com/zh/apis/x/posts/replies/list/)，$0.0020 / 一次） cURL ```bash curl -G "https://api.dataxapi.com/v1/x/posts/replies/list" \ --data-urlencode "id=2108198226532708497" \ -H "Authorization: Bearer $DATAXAPI_API_KEY" ``` Python ```python from dataxapi import DataXAPI dx = DataXAPI() # reads DATAXAPI_API_KEY result = dx.x.posts.replies.list(id="2108198226532708497") print(result["data"]["items"]) ``` TypeScript ```typescript import { DataXAPI } from "@dataxapi/sdk"; const dx = new DataXAPI(); // reads DATAXAPI_API_KEY const result = await dx.x.posts.replies.list({ id: "2108198226532708497" }); console.log(result.data.items); ``` 命令行 ```bash dataxapi call x.posts.replies.list id=2108198226532708497 ``` 录制的返回：取自 2026-10-12 19:11 UTC 录下的真实返回。名字、正文和编号是示例，链接也是，都不是录到的原文（页面不展示任何人的个人信息）；数字是原样。长列表做了截断。请求号、游标和余额是示意。 ```jsonc { "code": "ok", "request_id": "8d1c41ef-9476-4956-bef0-7cbcb9a72fe5", "data": { "items": [ { "id": "1852059004504439503", "url": "https://x.com/i/status/1852059004504439503", "type": "text", "text": "My honest take on weekend hiking, a month in.", "lang": "en", "published_at": "2026-10-08T14:10:01.000Z", "author": { "id": "185515870194801997", "username": "trail_and_lens", "name": "Maya Chen", "avatar_url": "https://pbs.twimg.com/profile_images/1839650090733403766/2e9uy1Sn_normal.jpg", "verified": true }, "metrics": { "likes": 14, "comments": 1, "favorites": 1, "views": 1419, "reposts": 3 }, "hashtags": [], "mentions": [ { "id": "120227434360127747", "username": "maya_makes", "name": "Northside Coffee" } ], "post_id": "1811536513306299064" }, // … 另有 5 条 ] }, "page": { "next_cursor": null, "cursor_expires_at": null, "has_more": false, "returned": 6 }, "warnings": [], "meta": { "capability": "x.posts.replies.list", "result_status": "complete", "fetched_at": "2026-10-12T19:11:03.000Z", "cached": false, "replayed": false, "unavailable": { "title": "platform", "metrics.shares": "platform", "is_pinned": "capability", "root_id": "platform", "author.followers": "capability", "images": "capability", "parent_id": "capability" }, "billing": { "amount": "0.002", "currency": "USD", "balance_after": "9.998" } } } ``` ## 示例花多少钱 取一次帖子本身，再翻三页回复 | 接口 | 调用次数 | 单价 | 小计 | | --- | --- | --- | --- | | [X 帖子详情](https://dataxapi.com/zh/apis/x/posts/get/) | 1 | $0.0003 | $0.0003 | | [X 帖子回复列表](https://dataxapi.com/zh/apis/x/posts/replies/list/) | 3 | $0.0020 | $0.0060 | | 合计 | 4 | | $0.0063 | 价格是生成本页时的现行价格。失败的调用、不存在的对象都不收费。 ## 在 AI 里做（MCP） 把 DataXAPI 接进你的 AI 客户端，只要接一次。之后直接说要什么，代理自己找到这些接口、看好价格，用你的余额调用。 这样说：用 DataXAPI 取 X 帖子 2108198226532708497 下面三页回复，按赞同、反对、提问分成三类。 代理发给 call 工具的参数： ```json { "capability": "x.posts.get", "id": "2108198226532708497", "idempotency_key": "" } ``` [了解 MCP 服务](https://dataxapi.com/zh/mcp/) ## 用 Skill 做 “评论与口碑”这个 Skill 把这件事做成了现成的任务，带花费上限。 “用户在这条内容下面夸什么、骂什么？” 它的示例运行：3 次调用，最多 $0.0060。 [评论与口碑](https://dataxapi.com/zh/skills/comments/) ## 要知道的 - 每页约 35 条，最多 60 条；游标一小时内有效。 - 每条回复的 `post_id` 是这段对话的主帖编号。它直接回复的是哪一条（`parent_id`）、回复者的粉丝数和图片地址都不返回。 - 顺序是 X 的相关度，不是时间；要按时间看，自己按 `published_at` 排。 - 不存在的帖子按来源出错返回，可以重试一次，不计费；回复返回空页照常计费。 ## 用到的接口 - [X 帖子详情](https://dataxapi.com/zh/apis/x/posts/get/) - [X 帖子回复列表](https://dataxapi.com/zh/apis/x/posts/replies/list/) ## X（推特）的更多内容 - [X（推特） API](https://dataxapi.com/zh/apis/x/) - [如何按关键词搜索 X（推特）帖子](https://dataxapi.com/zh/guides/x-search-posts-by-keyword/) - [如何获取 X（推特）各国热门话题](https://dataxapi.com/zh/guides/x-trends-by-country/) - [怎么查某个国家的 X 热搜？](https://dataxapi.com/zh/answers/how-to-get-x-trends-for-a-country/) --- 目录修订号 c48ce454ae6c4c65。 [网页版](https://dataxapi.com/zh/guides/x-post-replies/)