ae_kuaishou_web_user_id_from_share_link
### 用途 根据用户分享链接获取对应的用户 ID。 ### 输入 - `share_link`:必填。用户分享链接,例如短链接或用户资料分享链接。 ### 返回 成功响应中的 `data` 字段为从链接解析出的用户 ID。
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_kuaishou_web_user_id_from_share_link
### 用途 根据用户分享链接获取对应的用户 ID。 ### 输入 - `share_link`:必填。用户分享链接,例如短链接或用户资料分享链接。 ### 返回 成功响应中的 `data` 字段为从链接解析出的用户 ID。
ae_kuaishou_web_user_info
Look up a Kuaishou user's profile and account counts. Input: user_id is required and must be the user's eid as a string. Extract it from the profile URL: for https://www.kuaishou.com/profile/3xz63mn6fngqtiq, pass "3xz63mn6fngqtiq". A purely numeric uid is not supported; do not pass the full URL. Result: Returns user information directly. Profile data can include nickname, avatar URL, biography, and gender, along with follower, following, and received-like counts. In the documented response, data.userProfile.profile holds profile details and data.userProfile.ownerCount holds counts: fan means followers, follow means accounts followed, and like means received likes. Available fields may vary; this is not a video-list lookup. If request timeout is configurable, set it to more than 30 seconds. Requests are not guaranteed to succeed; if a request fails, try again later.
ae_kuaishou_web_video_comments
Fetch comments for a single video by providing its required photo_id. The optional pcursor is used for pagination: leave it empty or omit it on the first request, then use the pcursor returned in the response to fetch a subsequent page. The response provides comment data and may include status and request metadata.
ae_kuaishou_web_video_sub_comments
Retrieve replies (second-level comments) under one root comment on a Kuaishou video. Inputs: - photo_id (required string): ID of the video. - root_comment_id (required string): ID of the root comment whose replies you want to retrieve. - pcursor (optional string, default ""): Leave empty or omit for the first request. For subsequent pages, use the pcursor returned in the response with the same photo_id and root_comment_id. Returns: A response envelope containing the comment result in data, which may be null, along with response status information. The detailed fields within the comment result are not specified. This retrieves replies to the selected root comment, not all top-level comments on the video.
ae_wechat_get_account_services
Retrieve the custom menu and service entries shown at the bottom of a WeChat official account. Inputs: - request_body (required): An object containing username and optionally raw. - request_body.username (required): The account username. Supported forms are gh_…, gh_…@app for mini-program-linked accounts, and custom WeChat IDs such as nikejdi. Example: gh_363b924965e9. You can obtain this value from $.data.content.user_name in an article-detail response. - request_body.raw (optional boolean, default true): Use true for the original menu structure or false for a simplified parsed structure. Results: - raw=true: data contains the original menu structure. If the account has no configured custom menu, data is {}. - raw=false: $.data.biz_username contains the account username and $.data.menu contains the menu buttons, including names, types, submenus, and navigation targets where present. If no custom menu is configured, the menu list is []. This retrieves menu information; it does not activate the entries. Responses may be slow, so set the client request timeout to 30 seconds.
ae_wechat_get_article_ads
Retrieve embedded ads and ad-slot information from a WeChat Official Account article. Inputs: - request_body (required): an object containing the article URL and optional response mode. - request_body.url (required): an article URL on mp.weixin.qq.com, such as https://mp.weixin.qq.com/s/… or a /s link with query parameters including __biz. Maximum length: 2048 characters; it must match the URL format defined in the input schema. - request_body.raw (optional, default true): true returns full raw responses; false returns a flattened ad list. Results: - raw=false: data.count gives the number of ads merged from advertisement_info and ad_slot_data; data.ads contains the ad items. For an article without ads, data.ads is []. - raw=true: data.pages gives the number of captured response pages; data.results contains the full raw response for each page, including the advertisement_info / ad_slot_data fields that supply the flattened ad list. This retrieves ad information, not the article body. Responses may be slow; set the client request timeout to 30 seconds. The timeout is a client setting, not a tool parameter.
ae_wechat_get_article_stats_h5
## Purpose Get the total and elected comment counts for a WeChat Official Account article. This tool returns comment statistics only, not article text or individual comments. ## Inputs Pass an object named `request_body` containing: - `url` (required): A WeChat article URL on `mp.weixin.qq.com` using the `/s` path, such as `https://mp.weixin.qq.com/s/…`, or a supported long link with `__biz`. Maximum length: 2048 characters. - `raw` (optional, default `true`): Include raw comment counter fields when true; return counts without those extra fields when false. The top-level response structure is the same in both modes. - `comment_id` (optional): The article's comment ID as a numeric string of up to 32 characters, or `null`. If already available from article details at `$.data.content.comment_id`, supplying it can reduce parsing work. The URL is still required. ## Results and limitations - `$.data.comment_count`: Total comment count. - `$.data.elected_comment_count`: Elected comment count. - `$.data.read_num`, `$.data.like_count`, `$.data.old_like_count`, `$.data.share_count`, `$.data.collect_count` and `$.data.star_num`: Always `null` for this tool. Their names appear in `$.data.unavailable[]`. Unavailable does not mean zero. - With `raw=true`, `$.data.comment_raw` additionally contains `total_count`, `elected_total`, `comment_id` and `appmsgid`. Keep `comment_id` and `appmsgid` as strings to avoid loss of precision for large IDs. Set the client timeout to 30 seconds. If you need read counts or other unavailable engagement metrics, use a separate article-statistics capability; this tool does not automatically switch to one.
ae_wechat_get_collection_videos
## Purpose List videos in a known WeChat Channels collection, including video metadata, engagement counts, media references and pagination information. Each call returns one page, not necessarily the entire collection. ## Inputs Provide a required `request_body` object: - `topic_id` (required): digits-only collection ID as a string, at most 32 characters. Obtain it from a collection listing; for example, `"14616794565351966924"`. - `topic` (optional): collection name, up to 100 characters, to improve matching. - `username` (optional): creator finder username in `v2_<hexadecimal characters>@finder` format, up to 256 characters, to improve matching. - `topic_type` (optional): integer from 0 to 1000, default `16`. Use the type reported for the collection. - `last_buffer` (optional): base64 pagination cursor. Leave empty for the first page; pass the previous response's cursor unchanged for the next page. - `raw` (optional): defaults to `true` for the full original response. Set to `false` for simplified data, recommended for media handling. The optional text fields default to empty strings and also accept null. A name or username does not replace the required collection ID. ## Results and pagination With `raw=false`, `data` contains: - `topic_id` and `title`: collection identity and name. - `total_count`: total videos in the collection; `count`: items in the current page. - `videos[]`: video ID, creator username and nickname, title (which may be null), creation time, view/like/favorite/share/comment counts, and a `media` object. Media fields include `url`, `url_token`, `full_url`, `decode_key`, cover URL, duration, dimensions and media type. - `down_continue`: whether another page is available (`1` means yes); `last_buffer`: cursor for that page. With `raw=true`, videos are in `data.object[]`, the total is `data.totalCount`, and media fields are under `data.object[N].objectDesc.media[0]` as `url`, `urlToken` and `decodeKey`. Use `data.downContinueFlag` for collection pagination, not `continueFlag`. The cursor is `data.lastBuffer` (also documented as `data.last_buffer`); pass it as the next request's `last_buffer`. ## Important notes - Keep collection and video IDs as strings. Response IDs may exceed JavaScript's safe-integer range; use precision-safe JSON parsing and never convert them through JavaScript `Number`. - For media downloads, prefer the simplified `media.full_url`, which combines `url` and `url_token`. With raw data, concatenate `url` and `urlToken`. The bare URL may fail because of access-token requirements. - If a video is encrypted, decrypt it separately with the key from the same response. Keys change with each request. This tool does not download or decrypt files. - Responses may be slow; a 30-second client request timeout is recommended.
ae_wechat_get_live_detail
## Purpose Retrieve a specific WeChat Channels live room's anchor information, live status, audience counts, and replay information using its live ID. ## Inputs Pass a required `request_body` object containing: - `live_id` (required): A digits-only string, up to 32 characters, such as `2078950291205875194`. Obtain it from a live entry, shared live room, or live-history replay item. - `scene` (optional): Entry-scene integer from 0 to 1000; defaults to 5. - `finder_username` (optional): Anchor identifier in `v2_<hexadecimal characters>@finder` format, up to 256 characters. Providing it can improve matching. Defaults to an empty string; null is also allowed. - `raw` (optional): Defaults to `true`. Use `true` for the full original response or `false` for simplified fields. ## Returned information Live-room details are returned directly under `data`. - With `raw=false`, `data` uses snake_case fields: `live_id` and `object_id` identify the live room and associated video; `anchor_nickname`, `anchor_username`, and `anchor_head_url` describe the anchor; `live_cover_url` is the cover image. `live_status` indicates the live state (1 = live, 2 = ended; other values may occur), with text in `status_desc`. `online_count` and `max_online_count` represent current and peak online audiences. `replay_status`, `replay_url`, and `replay_watch_uv` describe replay availability, its address, and viewer count. - With `raw=true`, the full original structure includes `data.liveInfo` for live/replay details and `data.anchorContact` for anchor information. Examples include `data.liveInfo.liveId`, `liveStatus`, `onlineCnt`, `maxOnlineCount`, `replayUrl`, and `data.liveInfo.replayInfo.watchUv`. The associated object ID, status text, and replay URL may be empty. Online counts for ended streams are usually zero; a replay URL is not guaranteed. ## Usage notes Keep large IDs as strings throughout input and output handling. Use lossless JSON parsing rather than JavaScript `Number` to avoid rounding IDs beyond the safe-integer range. Responses can be slow; set the client request timeout to 30 seconds.
ae_wechat_get_related_articles
Retrieve the related articles displayed at the bottom of a WeChat Official Account article. Inputs: - request_body (required): Object containing the article URL and optional response format. - request_body.url (required): A WeChat article link on mp.weixin.qq.com, such as https://mp.weixin.qq.com/s/..., or a long /s?... link with __biz. Maximum length: 2048 characters. - request_body.raw (optional, default true): true returns the original response; false returns a simplified result. Results: - With raw=false, data.articles is the related article list, data.count is the number of articles, and data.ret is the return code (0 means success). - With raw=true, data.list is the original related article list, data.article_size is the number of articles, and data.base_resp.ret is the return code. Additional original-response metadata may also be present. - Articles without the related-articles feature return an empty list. This is not a general article search or an article-body retrieval tool. Responses may be slow. Set your client request timeout to 30 seconds.
ae_wechat_get_user_collections
## Purpose Get the collections listed on a WeChat Channels creator's homepage. This returns collection information, not the videos within each collection. ## Inputs Provide a `request_body` object containing: - `username` (required): the creator's finder username, in `v2_…@finder` format. It must be 10–256 characters and match `^v2_[0-9a-fA-F]+@finder$`; a display name cannot be used instead. - `raw` (optional boolean, default `true`): `true` returns the full raw structure; `false` returns a simplified structure. ## Results With `raw=false`, `data.username` identifies the creator, `data.count` is the total collection count, and `data.collections[]` contains: - `topic_id`: collection ID, usable for a separate collection-video lookup. - `name` and `desc`: collection name and description; the description may be empty. - `cover_url`: collection cover image URL. - `feed_count`: number of videos in the collection. - `topic_type`: collection type, usable with the collection ID for a separate video lookup. With `raw=true`, the corresponding fields are `data.collectionCount` and `data.collectionInfo[]`, with item fields `collectionTopicId`, `collectionName`, `collectionDesc`, `coverImgUrl`, `feedCount`, and `collectionTopicType`. `collectionTopicIdStr` provides a string alias for the ID. Raw data may also contain `lastUpdateTime`, `readCount`, and `categoryId`. `data.lastBuffer` is a collection-list pagination cursor, usually empty; this tool has no pagination input. ## Usage notes - Collection IDs may exceed JavaScript's safe-integer limit. Use lossless JSON parsing and preserve IDs as exact strings; use `collectionTopicIdStr` when available. Do not pass large IDs through JavaScript `Number`. - Responses can be slow. A 30-second client request timeout is recommended.
ae_youtube_get_channel_community_posts
## Retrieve YouTube channel community posts Get one page of community posts for a specified channel, with optional pagination. ### Inputs - **channel_id** (string, required): YouTube channel ID, such as `UCkRfArvrzheW2E7b6SVT7vQ`. You can obtain it from a channel URL of the form `https://www.youtube.com/channel/<channel_id>`. - **language_code** (string, optional; default `"zh-CN"`): Language code affecting returned text. - **country_code** (string, optional; default `"US"`): Country code, such as `US` or `JP`. - **continuation_token** (string, optional): Token from the previous response for retrieving the next page. Omit it for the first page. - **need_format** (boolean, optional; default `true`): Return cleaned, structured data when true; raw response data when false. ### Results With `need_format=true`, the response data contains: - **channel_id**: The channel ID. - **posts**: Community posts with `post_id`, `post_url`, author details (`author_name`, `author_channel_id`, `author_url`, `author_thumbnails`), `content_text`, `published_time` (relative time), `vote_count` (likes), `comment_count`, and `attachment`. Attachments may be an image, multiple images, video, poll, or null. - **continuation_token**: Token for the next page. - **has_more**: Whether more posts are available. - **total_count**: Number of posts on this page, not the total number on the channel. With `need_format=false`, the response data is raw and does not use the normalized structure described above. ### Pagination and limitations Start with `channel_id` and no continuation token. To continue, pass the returned `continuation_token` together with the same `channel_id`; in formatted results, stop when `has_more` is false. Some channels may not have a community/posts tab. This tool retrieves community posts, not the channel's video list or individual comments.
ae_youtube_get_channel_url
Find a YouTube channel's handle and links using its channel ID. Input: Required channel_id string, such as UCeu6U67OzJhV1KwBansH3Dg. Enter the channel ID, not an @username or channel URL. The documented result includes: - channel_id: Channel identifier. - handle: Channel username, such as CozyCraftYT. - title: Channel name. - channel_url: Standard channel link in /channel/ format. - vanity_url: User-friendly channel link in /@username format. Use this when you already have a channel ID and want to identify the channel or display a user-friendly link.
ae_youtube_get_post_comment_replies
Retrieve a page of replies to a YouTube post comment (second-level comments). Inputs: - continuation_token (required string): For the first page, use the comment's reply_continuation_token from a post-comments result. For subsequent pages, use the continuation_token returned by the previous replies result. - language_code (optional string): Language code; defaults to "zh-CN". - country_code (optional string): Country code; defaults to "US". - need_format (optional boolean): Whether to clean and format the data by extracting key content and removing redundant data; defaults to true. Returns: With need_format=true, the result data includes a comments list with comment_id, content, published_time, like_count, and author information. A continuation_token is provided when more replies are available. Set need_format=false to disable cleaning and formatting; the formatted field layout described here applies to need_format=true. Retrieve top-level post comments first to obtain the initial reply token. Each call retrieves a page, not necessarily the entire reply thread. Continue with the returned continuation_token while further pages are available.
ae_youtube_get_post_comments
## Purpose Retrieve comments on a YouTube community post, with pagination. This is for post comments, not video comments or replies to an individual comment. ## Inputs Provide at least one of the following: - `post_id` (string): The community post ID for an initial request. Extract it from a URL such as `https://www.youtube.com/post/UgkxiCSRfD6g7SPlWGPDa3vbP7aIsytXRkvy`. - `continuation_token` (string): Use the `comments_continuation_token` from post details to start, or the `continuation_token` from a previous comments response to retrieve the next page. Optional settings: - `language_code` (string): Language code; default `"zh-CN"`. - `country_code` (string): Country code; default `"US"`. - `need_format` (boolean): Clean and format the returned data; default `true`. The field layout below is documented for formatted results, not for unformatted results. ## Results With `need_format=true`, the result data includes: - `comments`: The comments on the requested page. Comment fields include `comment_id`, `content`, `published_time`, `like_count`, `like_count_a11y` (accessible like-count text), `reply_count`, and `reply_count_text`. - Each comment's `author`: Author information including `channel_id`, `display_name`, `channel_url`, `avatar_url`, `is_verified`, and `is_creator`. - Each comment's `reply_continuation_token`, where available: A token for retrieving that comment's replies with a separate reply-retrieval tool. - `continuation_token`: A token for requesting the next page of post comments when available. Pass the returned page token as `continuation_token` to continue reading comments. A reply token is not the next-page token for this list.
ae_youtube_get_post_detail
## YouTube community post details Retrieve the content, author information, attachments, and engagement counts for one YouTube community post. ### Inputs - **post_id** (string, required): The post ID, documented as starting with `Ugk`. Extract it from the `/post/` portion of a YouTube post URL; pass the ID, not the full URL. Example: `UgkxiCSRfD6g7SPlWGPDa3vbP7aIsytXRkvy`. - **language_code** (string, optional): Language code, such as `zh-CN` or `en-US`. Default: `zh-CN`. - **country_code** (string, optional): Country code, such as `US` or `JP`. Default: `US`. - **need_format** (boolean, optional): Whether to clean and format the returned data. Default: `true`. ### Returned data With `need_format=true`, the response data includes: - **post**: Post ID and URL; author name, channel ID, URL, and avatar thumbnails; text content; relative publication time; like count (`vote_count`); comment count; and attachment data. Attachments may be an image, multiple images, a video, or a poll, and may be `null`. - **channel_metadata**: Channel information such as name, description, and URL. - **comments_continuation_token**: A pagination token for subsequent comment retrieval, not the comments themselves. With `need_format=false`, the response data is returned in its raw form instead of the formatted structure.