ae_x_twitter_getpostsbyids
Get post details for multiple comma-separated post IDs.
AgentEarth Tool Catalog
Query social media profiles, posts, notes, comments and interaction data for account research, topic tracking and content performance analysis across platforms such as X (Twitter), RedNote (Xiaohongshu), Reddit and TikTok.
Tool List
ae_x_twitter_getpostsbyids
Get post details for multiple comma-separated post IDs.
ae_twitter_search_timeline
Search X (Twitter) for public conversations, accounts, media, or lists using the required keyword input. Optionally choose searchType: TOP for popular results, LATEST for latest results, MEDIA for media, PEOPLE for accounts, or LISTS for lists. No default result type is documented. Omit cursor or leave it empty for the first page; use a cursor returned by a previous response to request another page. The response contains the search payload in data, along with code, message, and recordTime. The detailed result structure and pagination field names are not specified, so available items and continuation information depend on the returned payload.
ae_tiktok_get_comment_replies
获取指定视频评论的回复数据。必须提供作品 ID(item_id)和评论 ID(comment_id);可选使用 cursor 分页游标与 count 设置数量。返回评论回复数据及请求状态等响应信息。
ae_dumplingai_get_tiktok_video
Get metadata, author details, and engagement stats for a TikTok video using DumplingAI.
ae_douyin_billboard_fetch_hot_account_trends_list
Retrieve trend data for a Douyin account's new likes, posts, comments, or shares. Inputs: - sec_uid (required string): The account's sec_id identifier. - option (optional integer): Metric to retrieve: 2 = new likes, 3 = new posts, 4 = new comments, 5 = new shares. Default: 2. - date_window (optional integer): Data-point interval in hours. Use 1 for hourly points or 24 for daily points. Default: 24. This sets granularity, not the overall date range. Returns: Account trend data for the selected metric in the response data payload. The payload's detailed structure is not specified and it may be null. The documented metric options do not include follower counts.
ae_chanmama_video_trend
View the latest trending topics across Douyin, Weibo, Zhihu, Bilibili, Baidu, Toutiao, and other platforms.
ae_douyin_billboard_fetch_hot_comment_word_list
Retrieve word-cloud weight data from comment analysis for a Douyin post. Input: aweme_id (required string), the ID of the target post. Returns: A response containing the comment-analysis payload in data, which may be null, plus response metadata. The analysis concerns word-cloud weights, not individual comment records. The payload's detailed structure and the method used to calculate weights are not specified, so weights should not automatically be treated as raw word counts.
ae_x_twitter_gettrendsbywoeid
Get trending topics for a location with the requested count.
ae_dumplingai_get_youtube_video
Get detailed metadata, statistics, and optional transcripts for a YouTube video using DumplingAI.
ae_xiaohongshu_topic_note_list
通过话题 ID 获取小红书话题下的笔记列表,支持按最热或最新排序以及游标分页。首次查询省略分页状态;后续页使用本工具上次返回的游标、搜索或会话状态,不能复用旧供应商的游标。
ae_douyin_fetch_douplus_promotable_item_list
## Purpose Retrieve the works available for promotion on a specified Douyin account's DOU+ promotion page, with item details and engagement metrics. This query does not start a promotion and is not an unrestricted list of all account posts. ## Inputs Pass a required `request_body` object containing: - `sec_uid` (required string): Identifier of the account whose promotable works you want to retrieve. - `count` (optional integer): Number of works to return; default is 30. - `target_id` (optional integer): Promotion goal used to filter eligible works; default is 31. Goals: 31 = likes and comments; 32 = followers; 33 = profile views; 37 = video plays; 40 = live-room popularity; 45 = comment-link clicks; 63 = show to followers; 78 = high-quality interaction. Keep 31 for a basic promotable-works query. - `aim_ids` (optional string): Advanced optimization objective codes, separated by commas. Normally omit this field so objectives are automatically matched to `target_id`. ## Result Returns response data containing the promotable works list, including information such as item IDs, titles, covers, and like, comment, and play counts. Eligibility is evaluated for the selected promotion goal. The detailed list structure is not specified, so available fields should be read from the actual response.
ae_xiaohongshu_pgy_fan_profile
根据用户ID获取小红书蒲公英创作者的粉丝画像及受众分布数据。参数请放在必填的 request_body 对象中。
ae_wechat_get_article_comments
## Purpose Get one page of featured comments on a WeChat Official Account article, with comment counts and pagination information. ## Inputs Supply a required `request_body` object: - `url` (required): WeChat article URL, for example `https://mp.weixin.qq.com/s/…`; maximum 2,048 characters. - `buffer` (optional, string or null; default `""`): Leave empty for the first page. For subsequent pages, copy the previous response's `data.buffer`; maximum 8,192 characters. - `raw` (optional, boolean; default `true`): `true` returns full raw comment objects; `false` returns a simplified comment view. - `comment_id` (optional, numeric string or null; maximum 32 digits): Article-level comment identifier. If already available from article details, supplying it can reduce latency. It is not an individual comment's `content_id`. ## Results `data.comments` contains the current page. The page summary includes `appmsgid` (article message ID), `elected_total` (featured comment count), `total_count` (total comment count), `comment_id` (article-level comment identifier), `buffer` (next-page cursor), and `has_more` (whether another page exists). In simplified mode, each comment includes its `content_id`, commenter nickname, text, likes, timestamp, IP region, featured/pinned flags, reply count (`reply_total`), and reply previews (`replies`). Raw mode preserves additional comment fields and represents reply information in `reply_new.reply_total_cnt` and `reply_new.reply_list`. Pass `data.buffer` back as `request_body.buffer` to continue; stop when `data.has_more` is `false`. The last-page cursor is an empty string. This is a single-page result, not a complete export of all comments or replies. Full second-level replies require separate retrieval using the individual comment's `content_id`. ## Handling notes Preserve large IDs as strings with precision-safe JSON parsing; do not convert them through JavaScript `Number`. Responses may be slow, so a 30-second client timeout is recommended. This is a client setting, not an additional tool input.
ae_tiktok_fetch_post_detail
获取单个短视频作品的数据。 输入: - `itemId`(必填,字符串):作品 ID。 - `region`(可选,字符串):内容地区代码,例如 `US`、`GB`、`JP` 或 `ID`;默认值为 `US`。 返回: - `data`:作品数据。 - 同时可能包含请求状态、请求参数、时间和请求标识等响应元数据。 提示:返回数据中的 `video.playAddr`、`video.downloadAddr` 以及 `video` 下各 `UrlList` 的首个链接可直接访问,无需 Cookie。请求这些链接时不要携带 `tt_chain_token`,否则可能返回 HTTP 403;`UrlList` 中的其他链接可能需要额外凭据。
ae_douyin_billboard_fetch_hot_total_high_topic_list
## Purpose Retrieve the Douyin ranking of topics with rising popularity, with pagination and optional keyword or content-category filters. ## Inputs Provide query options inside `request_body`: - `page`: Integer page number; default `1`. - `page_size`: Integer number of items per page; default `10`. - `date_window`: Time window in hours: `1`, `24`, `72`, or `168`, corresponding to the last hour, day, three days, or seven days. Default: `24`. - `keyword`: Keyword used to filter the ranking; default `""` means no keyword filter. - `tags`: Array of category filters; `[]` means all categories. Each entry uses `value` for the top-level category ID and optional `children` entries containing child-category `value` IDs. Omit `children` to include all subcategories under that top-level category. Obtain IDs from the content-tag lookup. ## Example ```json {"request_body":{"page":1,"page_size":10,"date_window":24,"keyword":"","tags":[]}} ``` ## Result Returns response data for the rising-popularity topic ranking matching the requested page, window, and filters. The response includes status and request metadata; individual topic fields within `data` are not defined by the documented response schema. This is a rising-topic ranking rather than a general popularity ranking.
ae_reddit_app_fetch_dynamic_search
## Search Reddit Search posts, communities/subreddits, comments, media (images, videos and GIFs), or users by keywords or a search expression. ### Inputs - **query** — Required, 1–512 characters. For post searches, ordinary keywords may match only some terms in titles or bodies. Use `rust AND borrow AND checker` when every term is required. - **search_type** — `post` (default), `community`, `comment`, `media`, or `people`. - **sort** — Optional: `RELEVANCE`, `HOT`, `TOP`, `NEW`, or `COMMENTS`. Only supported for post, comment and media searches; `COMMENTS` is post-only. Omit for community and people searches. - **time_range** — Optional: `all`, `year`, `month`, `week`, `day`, or `hour`. Only supported for post and media searches; omit for comment, community and people searches. These represent all time, past year, past month, past week, today and past hour, respectively. - **safe_search** — `unset` (default) or `strict`. - **allow_nsfw** — String `"0"` (default, disallow) or `"1"` (allow). - **language** — Preferred response language as an IETF language tag, such as `en-US` (default); 2–16 characters. - **after** — Pagination cursor, at most 4096 characters; defaults to an empty string for the first page. - **need_format** — Boolean, default `false`. Set to `true` for compact results and pagination information. ### Results and pagination Results are returned in `data`. With `need_format=false`, the original nested search structure is preserved. With `need_format=true`, `data` contains only: - **children** — A result list containing the corresponding post, comment, community or author content for the chosen search type. An empty list means no results. - **pageInfo** — Available pagination fields, including `endCursor`, `startCursor`, `hasNextPage` and `hasPreviousPage`. When `hasNextPage` is true, pass `pageInfo.endCursor` unchanged as `after`. Keep `query`, `search_type`, `sort` and other filters unchanged between pages. A null terminal cursor may be retained in pagination information. Compact results retain available IDs, titles, bodies, timestamps, statistics, authors, communities and media. Markdown is preferred for bodies. Redundant body representations, duplicate avatar sizes, presentation details, user interaction state and empty values are removed; meaningful `false` and `0` values are preserved. Fields vary by result type and availability. ### Post search expressions Use uppercase `AND`, `OR` and `NOT`, parentheses for grouping, and quotes for phrases. Supported field filters include `title:`, `selftext:`, `subreddit:`, `author:`, `flair:`, `site:`, `url:` and `self:` (`true` for text posts, `false` for non-text posts). Do not add spaces after field colons, and quote multi-word field values. Example: `title:"borrow checker" subreddit:rust`, optionally with `sort=TOP` or `sort=NEW`. `TOP` sorts by popularity and `NEW` by publication time; neither ensures that all query terms match. `time_range` limits result age without changing keyword matching. Minimal input: `{"query":"rust AND borrow","need_format":true}`.