Loading interactive reference

Fetching the local OpenAPI document and mounting the explorer.

API Reference

Social data and profile intelligence across Instagram, TikTok, YouTube, LinkedIn, Facebook, Douyin, Xiaohongshu (RED) and the Meta Ad Library, plus Instagram DM outreach. Every operation is one HTTPS call with a kb_live_… key, and every answer is { data, meta }. No SDK is needed.

Crawl jobs

Durable collection status and byte-bounded paid downloads. Polling and repeat downloads are free.

GET /v1/jobs/{job_id} Crawl job status

Owner-checked job status. Submission and polls cost zero. Crawler execution continues after disconnects. Successful completion charges once before results become available. Failed jobs never charge. A completed job waiting for funds returns 402; after top-up, settlement retries within 60 seconds.

GET /v1/jobs/{job_id}/results Crawl job results

Read paid results in responses bounded to 3 MiB. Send next_cursor to this route until complete is true. Download cursors differ from acquisition cursors. Each response may contain part of a target. Retries and repeated downloads never charge again. Before billing, the crawler confirms that each output unit fits 3 MiB in the selected raw or normalized mode. An oversized unit fails the job with result_item_too_large; results return 413 without a charge. Modes never switch automatically. If the crawler confirms that a retained snapshot is missing or corrupt after billing, the usage charge is reversed once and the job becomes failed. A later results read that finds the result again restores the job to ready and charges it again at the same price, or answers 402 when the balance cannot pay. Temporary store/network errors, invalid cursors and pending result preparation do not reverse charges. A failed reversal returns 503 billing_unavailable; retry the same result request. Results stay on the crawler and no unbounded aggregate crosses the portal.

POST /v1/jobs/{job_id}/cancel Cancel a crawl job

Stops a queued, running or payment-required job and marks it failed with error cancelled. A cancelled job is never charged, and its partial results are not delivered. Repeating the call returns the same answer. A job that already finished answers 409 job_finished.

GET /v1/jobs/platforms Supported platforms

The networks we store, each with how profiles are acquired (autonomous or authorized_required), its discovery modes, its content family, and what it supports: authoritative profiles, recent and deep content, comments, archive and media. Free.

Instagram

One folder: /v1/instagram/profile reads the tracked corpus (strictly DB); every other path is one Instagram operation, /v1/instagram/<op>, with cached=auto|cache_only|live and get_raw on each. usertags_feed, followers, following, story, likers, about, and highlight run on our signed-in account pool and are priced apart.

GET /v1/instagram/profile Profile

The canonical Instagram profile read model over the tracked corpus: followers, bio, category, engagement fields, and the AI enrichment block when present.

GET /v1/instagram/stored/profile/full Stored full profile

The stored full profile view in one call: profile (with AI enrichment fields when present), topContent, hashtags, collaborations, mentionedAccounts, brandCollabRate and a posting timeline. surface picks a smaller view: overlay or preview. Contact fields are not included; emails come from /v1/contacts/find-from-instagram/{fast,thorough}.

GET /v1/instagram/stored/profile/media Stored profile media

A profile block and its newest stored posts, each with every media item: ig_url (the Instagram URL), cdn_url (our mirrored copy, null until mirrored) and cdn_status. next_cursor is an opaque string, null on the last page; send it back as cursor for the next one. cdn_url is either a file URL or a signed /v1/media/{kind}/{id}/{slot} URL on this host that serves the same image without an API key. A signed URL works for at least 7 days, and a later read answers a fresh one.

GET /v1/instagram/stored/post Stored post

The stored record of one post: caption, likes, comments, posted_at, media URLs, content_type, product_type, size and flags.

GET /v1/instagram/stored/post/comments Stored post comments

The comments stored for one post, in the order they were collected, with stored_total, continuation (how far collection got) and next_cursor while has_more is true. A stored post with no collected comment page answers an empty list.

GET /v1/instagram/stored/post/media Stored post media

Every stored image or video item of one post, each with ig_url, cdn_url (our mirrored copy, null until mirrored) and cdn_status, plus the caption and counts. cdn_url is either a file URL or a signed /v1/media/{kind}/{id}/{slot} URL on this host that serves the same image without an API key. A signed URL works for at least 7 days, and a later read answers a fresh one.

GET /v1/instagram/stored/comments Stored recent comments

Stored comments across the posts of up to 100 accounts, newest first, each with the post it sits on. Name accounts by profile_id, the number /v1/instagram/profile returns. next_cursor is an opaque string while has_more is true; send it back as cursor for the next page.

GET /v1/instagram/web_profile_info web_profile_info

Wraps the logged-out profile reads: doc_id PolarisProfilePageContentQuery by ig_id when given, else the logged-out GraphQL profile by handle. Returns Profile under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/user_feed user_feed

Wraps GET /api/v1/feed/user/{username}/username/?count&max_id. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/highlights_tray highlights_tray

Wraps doc_id PolarisLoggedOutDesktopWWWProfileRootContentQuery (lox_highlights_connection, first 10). Returns Highlight[] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/post_comments post_comments

Wraps doc_id PolarisLoggedOutDesktopWWWPostCommentsPaginationQuery (first, after). With max_pages=1, returns Page[Comment] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/related_accounts related_accounts

Wraps doc_id PolarisAYMLFollowChainingListLoggedOutQuery (owner_id). Returns Account[] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/post post

Wraps GET /p/{shortcode}/. Returns Post under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/instagram/highlight highlight

Wraps GET /stories/highlights/{id}/. Returns Highlight under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/instagram/embed embed

Wraps GET /{username}/embed/. Returns Profile under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/instagram/location_info location_info

Wraps GET /explore/locations/{id}/ (preloaded PolarisExploreLocationsContainerQuery). Returns Location under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/instagram/location_nearby location_nearby

Wraps doc_id PolarisExploreLocationsNearbyTabRootQuery (location_id_str). Returns Location[] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/popular_search popular_search

Wraps doc_id PolarisLoggedOutPopularSearchPageQuery (media_count, related_keywords_count). Returns Keyword under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/popular_search_media popular_search_media

Wraps doc_id PolarisLoggedOutDesktopWWWPopularSearchPagePaginationQuery (media_count, after). With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/loggedout_profile loggedout_profile

Wraps doc_id PolarisLoggedOutDesktopWWWProfileRootContentQuery. Returns Profile under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/loggedout_timeline loggedout_timeline

Wraps doc_id PolarisLoggedOutDesktopWWWProfilePostsTabContentQuery (first, after). With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/usertags_feed usertags_feed (tagged media)

Wraps GET /api/v1/usertags/{id}/feed/?max_id&count. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/followers followers

Wraps GET /api/v1/friendships/{id}/followers/?count&max_id&rank_token. With max_pages=1, returns Page[Account] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/following following

Wraps GET /api/v1/friendships/{id}/following/?count&max_id&rank_token. With max_pages=1, returns Page[Account] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/story story

Wraps GET /api/v1/feed/user/{id}/story/. Returns Story[] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/likers likers

Wraps GET /api/v1/media/{id}/likers/. Returns Account[] under data; with get_raw=true returns the upstream JSON.

GET /v1/instagram/about about

Wraps POST bloks about_this_account. Returns About under data; with get_raw=true returns the upstream JSON.

Instagram DM

Message creators on Instagram and read their replies with two calls. We send from an Instagram account we run for your organization, so you never log in to Instagram and never pick an account. The DM playground shows the same conversations, which is the easiest place to see what happened.

POST /v1/instagram/dm/outreach Send outreach

Sends one Instagram message to a person. If you already have a conversation with them, the message continues it. Otherwise it starts one. We send from an Instagram account we run for your organization, so you never log in to Instagram.

GET /v1/instagram/dm/conversations/{conversation_id} Check conversation

Returns one conversation with every message, oldest first, including the person’s replies.

TikTok

One folder: /v1/tiktok library profiles, tracked videos and engagement analytics, the stored reads under /v1/tiktok/stored (strictly DB), and one wrapper per TikTok operation at /v1/tiktok/<op>.

GET /v1/tiktok/profile Profile

A TikTok account profile read model over the tracked corpus: nickname, bio, followers/following, video count, total hearts, verified/private flags, region, and secUid — plus the AI enrichment block when present.

GET /v1/tiktok/videos Videos

An account's tracked videos, newest first: caption, likes/comments/shares/plays, publish time, video/cover URLs, duration, and music. DB read over the tracked corpus — flat price for up to 100 videos a page. next_cursor is null on the last page; send it back as cursor for the next one. For a live walk of every video, use /v1/tiktok/item_list.

GET /v1/tiktok/stats Engagement analytics

Engagement analytics over an account's tracked videos: total/avg/max/median plays, avg likes/comments/shares, average engagement rate ((likes+comments+shares)/plays), 90-day posting cadence, and the top 3 videos by plays.

GET /v1/tiktok/stored/post Stored video

One stored video with its author: caption, counts, publish time, video and cover URLs, duration and sound.

GET /v1/tiktok/stored/post/comments Stored video comments

The comments stored for one video, in the order they were stored, with next_cursor while has_more is true. A stored video with no collected comment page answers an empty list.

GET /v1/tiktok/user_detail user_detail

The profile page's user detail: counts, bio, flags, avatar and bio links. extra carries sec_uid, heart_count and region. Returns Profile under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/tiktok/item_detail item_detail

One video from its public page: caption, counts, sound, cover and author. Returns Post under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/tiktok/item_transcript item_transcript

The video's native caption track as text. This reads captions TikTok already has; it does not transcribe. A video with no native caption answers 502 and is not billed. Returns Transcript under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/tiktok/item_list item_list

The account's videos, newest first. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/challenge_item_list challenge_item_list

A hashtag's videos. TikTok pads the feed with videos that lack the tag, so only videos that carry it are answered; get_raw=true returns every native item. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/challenge_detail challenge_detail

A hashtag's own page: its id, title, description, and how many videos and views TikTok counts for it. Returns Challenge under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/music_item_list music_item_list

The videos that use one sound, with each video's author. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/search_item_full search_item_full

TikTok's video search, ranked by relevance across years. Search a brand name to find the videos that talk about it. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/search_user_full search_user_full

TikTok's account search, such as a brand's regional handles. With max_pages=1, returns Page[Account] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/comment_list comment_list

A video's top-level comments with author, likes and reply count. With max_pages=1, returns Page[Comment] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/related_item_list related_item_list

The videos TikTok ranks as similar to one video. TikTok picks them by similar content, so they find creators in the same niche; they do not show duets, stitches or collaborations. Returns Post[] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/followers followers

The accounts that follow this account, as TikTok serves them without a login. With max_pages=1, returns Page[Account] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/following following

The accounts this account follows. With max_pages=1, returns Page[Account] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/user_playlist user_playlist

The playlists (series) a creator curates, with each playlist’s video count. An account with no playlists answers an empty list. Returns Playlist[] under data; with get_raw=true returns the upstream JSON.

GET /v1/tiktok/mentioned-in Videos mentioning an account

Videos by other creators that @mention an account, newest first. Point it at a brand to list creators who posted about it.

TikTok Shop

Stored TikTok Shop data under /v1/tiktok_shop/…: product and shop search, the category tree, one product, shop or creator, and the videos and affiliate creators that sell them, with daily units and GMV. Strictly DB: an id the store does not hold is a 404 and is not billed. Ids are strings because TikTok ids have 19 digits.

GET /v1/tiktok_shop/products Search TikTok Shop products

Products that sold or have a selling video, searched by title, brand or shop name and filtered by market, category subtree, shop, price, 30-day units, lifetime units and status. The default sort is relevance with q, else units_30d; ties go to lifetime units sold. total is exact up to 10,000; past that it is a floor and total_exact is false.

GET /v1/tiktok_shop/shops Search TikTok Shop shops

Every stored shop of a market, searched by shop name. The default sort is relevance with q, else gmv_30d.

GET /v1/tiktok_shop/categories TikTok Shop category tree

The stored category tree of a market, children sorted by name. With parent_id, only the subtree under that category. Reference data for the category_id filter, so it is free.

GET /v1/tiktok_shop/product TikTok Shop product

One product with its SKUs, category path, rating histogram, shop, video and creator counts, top selling videos by plays, and daily units and GMV for days days. If the sales store does not answer, the product still returns with sales.available false and sales.error naming the cause.

GET /v1/tiktok_shop/product/videos TikTok Shop product videos

Videos that sell the product: anchored on the video or listed on the product page. Each row names its creator and the paid or affiliate label TikTok shows.

GET /v1/tiktok_shop/product/creators TikTok Shop product creators

Creators whose videos sell the product, with their video count and plays.

GET /v1/tiktok_shop/shop TikTok Shop shop

One shop with its business details, product, video and creator counts, top products by 30-day units, and daily units and GMV summed over its products for days days. If the sales store does not answer, the shop still returns with sales.available false.

GET /v1/tiktok_shop/shop/products TikTok Shop shop products

Every listed product of the shop, sold or not. An empty sort value sorts last.

GET /v1/tiktok_shop/shop/creators TikTok Shop shop creators

The shop's affiliate creators: every creator whose videos sell one of its products, with videos, distinct products and plays. Each video counts once.

GET /v1/tiktok_shop/shop/videos TikTok Shop shop videos

Videos that sell the shop's products, one row per video with the products it sells.

GET /v1/tiktok_shop/creator TikTok Shop creator

The shop side of a creator: their TikTok account once resolved, video, product and shop counts, the shops and products their videos sell (most videos first), and their top selling videos by plays. It reads stored TikTok Shop data only and never crawls, so read any other TikTok account with /v1/tiktok/user_detail.

GET /v1/tiktok_shop/creator/videos TikTok Shop creator videos

The creator's selling videos, each with the products it sells.

YouTube

One folder: /v1/youtube library profiles, tracked videos and view analytics, the stored reads under /v1/youtube/stored (strictly DB), and one wrapper per YouTube operation at /v1/youtube/<op>.

GET /v1/youtube/profile Channel

A YouTube channel profile read model over the tracked corpus: name, bio, subscriber/video/view counts, country, verified flag, keywords, and avatar/banner URLs — plus the AI enrichment block when present.

GET /v1/youtube/videos Videos

A channel's tracked videos, newest first, with video id, title, description, views, likes, comments, publish time, duration, Shorts/live flags, category, keywords, and thumbnail when observed. Videos with no publish time come last. DB read over the tracked corpus — flat price for up to 100 videos a page. next_cursor is null on the last page; send it back as cursor for the next one. For a live walk of a channel's tab, use /v1/youtube/channel_videos.

GET /v1/youtube/stats View analytics

Performance analytics over a channel's tracked videos: tracked count, total/avg/max/median views, average duration, total likes/comments, average engagement rate, and the top 5 videos by views. Numbers describe the tracked corpus (tracked_videos), not the channel's full history; legacy sparse fields remain null.

GET /v1/youtube/stored/post Stored video

One stored video with its channel: title, description, counts, publish time, duration, the paid partnership flag and its collaborators.

GET /v1/youtube/stored/post/comments Stored video comments

The comments stored for one video, in the order they were stored, with next_cursor while has_more is true. A stored video with no collected comment page answers an empty list.

GET /v1/youtube/channel channel

The channel's About panel: subscribers, videos, links and join date. extra carries banner_url, view_count, country, joined_date and keywords. Returns Profile under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/channel_videos channel_videos

The channel's Videos tab. A listed video carries its title, views and a relative date (extra.published_text); include_extras=true reads each watch page for the exact date, likes and description, and paid_only=true keeps the videos that disclose a paid promotion. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/channel_shorts channel_shorts

The channel's Shorts tab. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/channel_lives channel_lives

The channel's Live tab, newest first. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/channel_playlists channel_playlists

The playlists the channel shows. With max_pages=1, returns Page[YouTubePlaylist] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/channel_community_posts channel_community_posts

The channel's community posts. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/community_post community_post

One community post from its public page. Returns Post under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/youtube/playlist playlist

A playlist's videos in playlist order. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/search search

YouTube's search: videos, channels and playlists, ranked. With max_pages=1, returns Page[YouTubeSearchResult] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/search_hashtag search_hashtag

The videos YouTube finds for a hashtag. With max_pages=1, returns Page[YouTubeSearchResult] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/search_suggestions search_suggestions

YouTube's autocomplete for a query. Returns YouTubeSuggestion[] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/video video

One video from its watch page: date, counts, description, duration and the paid promotion flag. extra.title is the title. A removed or unplayable video answers 404. Returns Post under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/video_comments video_comments

A video's top-level comments. A comment with replies carries extra.replies_token for /v1/youtube/video_comment_replies. With max_pages=1, returns Page[Comment] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/video_comment_replies video_comment_replies

The replies to one comment. With max_pages=1, returns Page[Comment] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/video_transcript video_transcript

The video's caption track in one language as timed segments and text. This reads captions YouTube already has; it does not transcribe. A video with no track in that language answers 404 no_captions and is not billed. Returns YouTubeTranscript under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/video_sponsors video_sponsors

Sponsor names the description or caption track states, with the phrase that names each. This is our heuristic, not YouTube's own list. Returns YouTubeSponsors under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/video_thumbnails video_thumbnails

Every thumbnail size the watch page names. Returns YouTubeThumbnail[] under data; with get_raw=true returns the upstream JSON.

GET /v1/youtube/profile_full profile_full

The channel, its newest videos read from their watch pages, and engagement, cadence and format figures over them. posts is how many videos to read, 1 to 100, default 25. Returns YouTubeProfileFull under data; with get_raw=true returns the upstream JSON.

LinkedIn

Public LinkedIn reads under /v1/linkedin/<op> (company pages, members, posts, jobs, job search and the LinkedIn Ad Library) and stored reads under /v1/linkedin/stored/….

GET /v1/linkedin/stored/company Stored company

One stored company page with its size, industry, headquarters and followers.

GET /v1/linkedin/stored/company/posts Stored company posts

The stored posts of one company. Newest first, with an opaque next_cursor while has_more is true; send it back as cursor.

GET /v1/linkedin/stored/company/jobs Stored company jobs

The stored jobs of one company, by publish time. Newest first, with an opaque next_cursor while has_more is true; send it back as cursor.

GET /v1/linkedin/stored/company/ads Stored company ads

The stored Ad Library ads of one advertiser, by first impression day. Newest first, with an opaque next_cursor while has_more is true; send it back as cursor. A numeric q reads the advertiser id even when its company page is not stored.

GET /v1/linkedin/stored/member Stored member

One stored member profile.

GET /v1/linkedin/stored/member/posts Stored member posts

The stored posts of one member. Newest first, with an opaque next_cursor while has_more is true; send it back as cursor.

GET /v1/linkedin/stored/post Stored post

One stored post with its author and counts.

GET /v1/linkedin/stored/post/comments Stored post comments

The comments stored for one post, in the order they were stored, with next_cursor while has_more is true.

GET /v1/linkedin/stored/job Stored job

One stored job posting.

GET /v1/linkedin/stored/ad Stored ad

One stored Ad Library ad.

GET /v1/linkedin/profile profile

A member's guest profile: name, headline, about, followers and connections. extra carries headline, location, connections and the current company. Returns Profile under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/company company

A company page: name, about, industry, headquarters, website and followers. extra.company carries its size, type, founding year, specialties and job count. Returns Profile under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/company_full company_full

The company, every post its guest page shows, newest first, and posting cadence, engagement and format figures over them. Returns LinkedInCompanyFull under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/company_job_count company_job_count

How many jobs a company has open worldwide; count_exact is false for a rounded figure. Returns LinkedInJobCount under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/company_jobs company_jobs

A company's open jobs, 10 a page, up to 100 pages. since needs sort_by=recent. Returns Page[LinkedInJob] under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/search_jobs search_jobs

LinkedIn's guest job search, 10 a page, up to 100 pages. since needs sort_by=recent. Returns Page[LinkedInJob] under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/job job

One job posting with its description, criteria, applicants and company. A closed job answers with extra.closed true. Returns LinkedInJob under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/post post

One post with its author, reactions, comments and reposts. extra.comments holds the first comments the guest page shows; LinkedIn pages no further comments to a signed-out reader. Returns Post under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/ad ad

One Ad Library ad: advertiser, payer, format, copy, run dates, impressions and targeting. Returns LinkedInAd under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/ads_search ads_search

The Ad Library's search for a keyword, 24 ads a page, each read from its detail page. Returns Page[LinkedInAd] under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

GET /v1/linkedin/company_ads company_ads

The ads one advertiser ran, 24 a page, each read from its detail page. Returns Page[LinkedInAd] under data. LinkedIn is read as a signed-out guest and every upstream document is an HTML page, so get_raw is not available.

Facebook

Public Facebook reads, one path per operation at /v1/facebook/<op> on the Instagram grammar (cached, get_raw, and since, max_pages and cursor on paged operations), and the stored reads under /v1/facebook/stored. The Meta Ad Library reads are under /v1/meta_ads. Transient challenge failures never become empty successes.

GET /v1/facebook/stored/profile Stored page

One stored page: name, category, followers, links and when it was last read, with metric_precision naming the counts Facebook rounds.

GET /v1/facebook/stored/profile/posts Stored page posts

The posts stored for one page, newest first, with an opaque next_cursor while more are stored.

GET /v1/facebook/stored/post Stored post

One stored post with its page and group: caption, counts, time, media paths and metric_precision.

GET /v1/facebook/stored/post/comments Stored post comments

The comments stored for one post, in the order they were stored, with next_cursor while has_more is true. A stored post with no collected comment page answers an empty list.

GET /v1/facebook/profile profile

A page's public profile: name, category, followers and links. extra carries likes, website, location, delegate_page_id and, with get_business_hours=true, business_hours. Returns Profile under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/profile_photos profile_photos

The photos a page shows, 8 a page. Returns Page[Post] under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/profile_reels profile_reels

A page's reels, 10 a page, with views, likes and comments. Returns Page[Post] under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/profile_events profile_events

The events a page hosts, 8 a page. Returns Page[FacebookEvent] under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/group_posts group_posts

A public group and the posts its logged-out discussion page shows, usually only the newest one. Facebook gives a logged-out reader no further page. Returns FacebookGroupPosts under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/post post

One post, reel, video or photo from its public page. Send its URL to /v1/facebook/post_comments for its comments. Returns Post under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/post_comments post_comments

A post's top-level comments, 10 a page. Send a comment's id as feedback_id, with this post's URL as url, to /v1/facebook/post_comment_replies for its replies. With max_pages=1, returns Page[Comment] under data; with get_raw=true returns the upstream JSON.

GET /v1/facebook/post_comment_replies post_comment_replies

The replies to one comment, 10 a page, read on the page of the post the comment is on. A call by the comment’s numeric id costs the crawler one more request, which reads its current feedback id and token. A comment with no replies answers an empty page. With max_pages=1, returns Page[Comment] under data; with get_raw=true returns the upstream JSON.

GET /v1/facebook/post_transcript post_transcript

A video's own caption track as Facebook serves it to a logged-out viewer. This reads captions Facebook already has; it does not transcribe. A video with no track answers 404 and is not billed. Returns FacebookTranscript under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/event_details event_details

One event by id or URL. Facebook answers an unknown event id with a generic page that reads as a block, so take ids from /v1/facebook/profile_events or /v1/facebook/events_search. Returns FacebookEvent under data; the upstream document is an HTML page, so get_raw is not available.

GET /v1/facebook/events_search events_search

Facebook's public event search for a query. While the residential proxy pool is out, the search answers one page of about 8 events: a search with more results answers next_cursor: null and stop_reason: "cursor_stalled", and a call that sends a cursor answers 400. Facebook answers some queries with a normal-looking empty page, so a first page with no events carries meta.possibly_shadow_empty: true and may be false. With max_pages=1, returns Page[FacebookEvent] under data; with get_raw=true returns the upstream JSON.

Xiaohongshu

Xiaohongshu (RED): logged-out explore, creator and note reads at /v1/xiaohongshu/<op>, and stored notes and creators under /v1/xiaohongshu/stored/… (strictly DB).

GET /v1/xiaohongshu/stored/note Stored note

The canonical RED note read model over the tracked corpus, including engagement, ordered source/mirrored media, author identity, and partial/full state. Strictly DB: a miss is 404 and never triggers acquisition. Read a fresh note with /v1/xiaohongshu/note_detail.

GET /v1/xiaohongshu/stored/profile Stored creator

A stored Xiaohongshu creator: name, red id, bio, avatar, follower and following counts with their precision, whether a creator page (is_authoritative) or only a note card named it, and when we last read it. Strictly DB: a creator we never stored is a 404, never a crawl. Read a fresh one with /v1/xiaohongshu/user_profile.

GET /v1/xiaohongshu/stored/profile/notes Stored notes

One page of a stored creator's notes, newest first, a note read only from an explore card (is_partial) last. Pass next_cursor back as cursor for the next page; has_more is true exactly when it is set. Flat price per page. Strictly DB: a creator we never stored is a 404, never a crawl.

GET /v1/xiaohongshu/explore explore

One explore page of about 30 notes, each a feed card with its title, like count, cover and author (extra.is_partial is true). Each first page is a fresh draw, and next_cursor reads the pages after it. Returns Page[Post] under data.

GET /v1/xiaohongshu/user_profile user_profile

A creator's page: name, bio, avatar, follower and following counts. extra carries red_id, ip_location, gender, professions, likes_and_collects and the grid's note cards, which Xiaohongshu serves without note ids. Rounded counts are named in extra.approximate_metrics. Returns Profile under data.

GET /v1/xiaohongshu/note_detail note_detail

One note: title, text, posted time, like, comment, save and share counts, tags, mentions, images or video and author. Call /v1/xiaohongshu/explore first and pass a note_id from its answer: Xiaohongshu serves a note only with the token its feed issued, and the crawler keeps that token from the explore read. Any other id, the example included, fails at once with 400 invalid_request and is not billed. Returns Post under data.

Douyin

Douyin (the Chinese TikTok): one wrapper per Douyin operation at /v1/douyin/<op> (user_profile, aweme_post, aweme_detail, comment_list, aweme_related) on the Instagram grammar, and stored profiles and posts by sec_uid under /v1/douyin/stored/… (strictly DB).

GET /v1/douyin/stored/profile Douyin stored profile

A stored Douyin profile: name, bio, follower, following, post and like counts, and when we last read it. Strictly DB: a profile we never stored is a 404, never a crawl. Read a fresh one with /v1/douyin/user_profile.

GET /v1/douyin/stored/profile/posts Douyin stored posts

One page of a stored Douyin profile’s posts, newest first, with counts, caption, cover and video. Pass next_cursor back as cursor for the next page; has_more is true exactly when it is set. Flat price per page. Strictly DB: a profile we never stored is a 404, never a crawl.

GET /v1/douyin/stored/post/comments Douyin stored comments

The comments stored for one Douyin video, in the order they were stored, with each commenter’s uid, sec_uid, nickname and region (ip_label). Pass next_cursor back as cursor for the next page; has_more is true exactly when it is set. Flat price per page. Strictly DB: comments collect through /v1/douyin/comment_list, and a video with neither a stored read nor a stored comment is a 404, never a crawl.

GET /v1/douyin/user_profile user_profile

A profile: name, bio, follower, following, post and like counts, verification and avatar. extra carries sec_uid, total_favorited and favoriting_count. Returns Profile under data; with get_raw=true returns the upstream JSON.

GET /v1/douyin/aweme_post aweme_post

The account's newest posts, up to 42. Douyin withholds every later page from a logged-out reader, so a longer feed stops with stop_reason cursor_stalled and no next_cursor. With max_pages=1, returns Page[Post] under data; with get_raw=true returns the upstream JSON.

GET /v1/douyin/aweme_detail aweme_detail

One video or image post: caption, counts, hashtags, sound, location and author. An unknown or removed video answers 404 and is not billed. Returns Post under data; with get_raw=true returns the upstream JSON.

GET /v1/douyin/comment_list comment_list

A video's top-level comments, hottest first, with each commenter's region in extra.ip_label. With max_pages=1, returns Page[Comment] under data; with get_raw=true returns the upstream JSON.

GET /v1/douyin/aweme_related aweme_related

The videos Douyin ranks as related to one video, with their authors. Douyin picks them by similar content, so they find creators in the same niche; they are not collaborations. Returns Post[] under data; with get_raw=true returns the upstream JSON.

Meta Ads

Meta Ad Library reads at /v1/meta_ads/<op> on the Instagram grammar (cached, get_raw, and max_pages and cursor on paged operations), where an ad answers as a Post and an advertiser as an Account, and stored reads under /v1/meta_ads/stored/… (strictly DB): full-text ad search with structured filters, advertiser profiles and ads, and aggregate ad-strategy insights.

GET /v1/meta_ads/stored/advertisers Search advertisers

The tracked advertiser directory: a prefix match on Page name or handle, sorted by active ads, total ads or Page likes. Without query it lists every tracked advertiser. total counts every match, and next_cursor ({ offset }) with has_more names the next page. Use a pageId with /v1/meta_ads/stored/advertiser, /v1/meta_ads/stored/advertiser/ads, /v1/meta_ads/stored/insights or /v1/meta_ads/stored/search.

POST /v1/meta_ads/stored/advertisers/lookup Look up advertisers

Many tracked advertisers in one call, for a watchlist: each row adds the latest ad start (latestAdStartedAt), the newest ingest time (ingestCursor), and newAdCount, the ads ingested after that page's newSince cursor. Store ingestCursor as the next newSince. Billed at the advertiser price per advertiser returned; an id the corpus does not hold is left out and not billed.

GET /v1/meta_ads/stored/advertiser/ads Stored advertiser's ads

One tracked advertiser's stored ads as Posts, running ads first and then the newest start. Pass next_cursor back as cursor while has_more is true; active_only=true lists only running ads. The answer also carries the advertiser as an Account.

GET /v1/meta_ads/company_ads company_ads

The ads one advertiser (a Facebook Page) runs or ran, up to 30 a page in the Ad Library's order. Find a page_id with /v1/meta_ads/search_companies. Returns Page[Post] under data, read from the Ad Library as a signed-out guest.

GET /v1/meta_ads/search_ads search_ads

The Ad Library keyword search over ad copy, up to 30 ads a page, across every advertiser. Returns Page[Post] under data, read from the Ad Library as a signed-out guest.

GET /v1/meta_ads/search_companies search_companies

The Pages the Ad Library typeahead suggests for a name, each with its page_id as id; category, Page likes and the Page URL sit under extra. Returns Account[] under data, read from the Ad Library as a signed-out guest.

GET /v1/meta_ads/ad ad

One ad by its Ad Library id, read from its advertiser's library page: copy, creatives, run dates, CTA, platforms and countries. An id the library does not answer is a 404. Returns Post under data, read from the Ad Library as a signed-out guest.

POST /v1/meta_ads/stored/search Search stored ads

Full-text search over tracked Meta Ad Library ads with structured filters: advertiser name, page id, CTA type, link domain, publisher platform, media type, active-only, minimum days active, and start date. Sortable by start date, longevity (daysActive), or A/B-test breadth (collation).

GET /v1/meta_ads/stored/advertiser Advertiser

A tracked advertiser (Meta Page): page name/username, category, verification, Instagram link, page likes, and total/active ad counts — the discovery pivot for per-advertiser monitoring.

GET /v1/meta_ads/stored/insights Advertiser insights

Aggregate ad-strategy insights for one advertiser: total/active ad counts, longevity (active ads age dynamically from start_date; ended ads use their recorded lifetime), display-format / CTA / publisher-platform distributions, top link domains, a 6-month launch series, and the 5 longest-running active ads.

Crawl

The scrape-any-URL product (POST /v1/crawl/web-scraping-api) plus portal-owned presets for non-core networks (/v1/crawl/{network}/{resource}). Presets compile to one network-blind crawler door, and every fetch runs through residential proxies, with SSRF, redirect, and size boundaries enforced server-side.

POST /v1/crawl/web-scraping-api Scrape any public URL

Fetch any public url and get the document back. A call with no geo fetches from our own servers, and a geo target sends it through our residential proxy pool. Optional extract maps output fields to dot-paths in the parsed JSON. Optional geo targets the egress: session (8 lowercase alphanumerics — reuse the same value across calls to keep one exit IP), lifetime (minutes, 3–1440, requires session), and type (residential/mobile/datacenter). A geo target is either honoured or refused, never silently substituted — an egress we cannot serve returns 400 and is not charged. Optional spoof decides how the request PRESENTS itself: render executes the target page JavaScript in a real de-automated browser (for surfaces that serve no content to a plain HTTP client), warm_url establishes the target cookies first in the same browser context, wait_for is a CSS selector proving the page produced its content (echoed back as wait_matched), plus headers, user_agent, locale, accept_consent, timeout_ms. The /v1/search/* SERP endpoints are exactly this door with a search-engine URL and a spoof block. Only public destinations; http/https GET only; no credentials or arbitrary headers pass through; redirects, size, and time are bounded.

GET /v1/crawl/hackernews/profile Hacker News user

A Hacker News user — karma, account age, about text, and their submissions. A username that does not exist returns no fields and is not billed.

GET /v1/crawl/hackernews/story Hacker News item

One Hacker News item — story, comment, poll, or job — by its numeric id. An id with no item returns no fields and is not billed.

GET /v1/crawl/hackernews/story/comments Hacker News comment tree

An item's full nested comment tree, as served by the Algolia items API.

GET /v1/crawl/hackernews/search Hacker News search

Full-text Hacker News search over stories, via the Algolia search API. A search with no hits is not billed.

GET /v1/crawl/bluesky/profile Bluesky profile

A Bluesky account — handle, display name, description, and follower/post counts.

GET /v1/crawl/bluesky/post Bluesky post thread

One Bluesky post and its reply thread, by AT-URI.

GET /v1/crawl/bluesky/user/posts Bluesky author feed

A Bluesky account's recent posts (author feed).

GET /v1/crawl/linktree/page Linktree page

A Linktree creator page — the document, with the page's links available via HTML extraction.

GET /v1/crawl/x/tweet X (Twitter) tweet

One tweet by numeric id — text, author, like count, and created date — via the public syndication CDN.

GET /v1/crawl/threads/profile Threads profile

A Threads profile page — the document, with bio and follower data embedded in the HTML. A page with no profile data is not billed.

GET /v1/crawl/google/play-app Google Play app

A Google Play store listing by package name — the document, with JSON-LD app metadata embedded.

GET /v1/crawl/google/news Google News search

Google News search results for a query, as the public RSS feed (XML). A search with no results is not billed.

GET /v1/crawl/biolink/page Bio Link page

A bio.link creator page — the document, with the page's links available via HTML extraction.

GET /v1/crawl/komi/page Komi page

A Komi creator profile from the JSON API the Komi page loads: name, bio, avatar, account status and social links. A username with no Komi page returns an error and is not billed.

GET /v1/crawl/snipfeed/page Snipfeed page

A Snipfeed creator page — the document, with the page's links available via HTML extraction.

GET /v1/crawl/taplink/page Taplink page

A Taplink creator page — the document, with the page's links available via HTML extraction. A username with no Taplink page is not billed.

GET /v1/crawl/carrd/page Carrd site

A Carrd one-page site by its subdomain — the document, with links available via HTML extraction.

Web Search

Live web search against a named engine — /v1/search/google and /v1/search/bing — returning organic results in the DataForSEO serp/{engine}/organic/live/regular response shape, with pagination via depth. The engine you ask for is the engine that answers: we never substitute another index behind the name. A blocked engine returns 502 upstream_error (and is never billed), never an empty result list, so "no results" always means the engine genuinely had none.

GET /v1/search/google Google web search

Live Google organic SERP in the DataForSEO serp/google/organic/live/regular response shape (data.tasks[].result[].items), with pagination: depth > 10 fans the query across result pages and each item carries its page.

GET /v1/search/bing Bing web search

Live Bing organic SERP in the same DataForSEO serp/{engine}/organic/live/regular response shape as /v1/search/google (with se_domain: bing.com), priced lower because Bing tolerates our egress better.

GET /v1/search/locations Supported search markets

Every market /v1/search/* accepts, as { location_code, location_name, country_iso_code }. Google searches the market you name; Bing searches in the country it sees the request come from. Free and unbilled — you should never pay to discover which values the API accepts. location_code is a Google Ads geo-target criterion id, so this is the only way to map a country to one; the field names mirror DataForSEO, whose reference endpoint this matches.

GET /v1/search/languages Supported search languages

Every language /v1/search/* can search, as { language_code, language_name }. Free and unbilled. A code outside this list is rejected rather than forwarded to the engine, so a caller never receives results for a market nobody serves.

Creator Search

Search the stored creator index under /v1/creators/…: creators by brand collab, profiles by audience filters, posts by caption and collab, and profile lookups by handle or id, across networks. Every call reads stored data and never crawls. Rows carry no contact data; find contacts with /v1/contacts/find.

POST /v1/creators/search Find creators by brand collabs

Creators who collaborated with a brand, newest collab first. brandHandles and the collabSince window apply to the same collab, so a creator whose only glossier collab is a year old does not match "glossier in the last week". Add free text, followers, country, niche and engagement filters, sort by last_collab, followers, relevance or engagement, and ask for facets. Each row carries its collab edges.

POST /v1/creators/profiles/search Search profiles

Profile search with the full filter set: followers, niche, country, language, gender, engagement and posting ranges. Rows carry the stored profile, AI enrichment fields, engagement statistics and recent post previews; set omit: ["previewPosts"] for smaller pages, or name the row fields you read in fields. Up to 200 rows per page.

POST /v1/creators/content/search Search posts

Post search over captions with platform, niche, country, date, likes and paid partnership filters, and brand collab retrieval through collabBrands or collabOf. Page past offset 10000 with searchAfter. With mode: "preview", the answer is the newest posts per profile in itemsByProfileId.

POST /v1/creators/mget Creator rows by handle

The creator search rows for up to 200 handles, with their collab edges. Without platform, a handle that exists on several platforms returns one row per platform. A call that finds none of the handles costs $0.

GET /v1/creators/profiles Instagram profiles by handle

Stored Instagram profiles for up to 200 handles: bio, counts, niche, country, language and scrape times. A handle we never crawled is missing from the list, and a call that finds none of its handles costs $0.

GET /v1/creators/profile Creator profile

One creator by the profile id that search, autocomplete and mget return. surface=full adds top content, hashtags, collaborations, mentioned accounts, brand collab rate and the post timeline to the profile; overlay and preview are smaller cuts.

GET /v1/creators/autocomplete Autocomplete creators

Handle suggestions for a search box: accounts whose handle starts with q, largest first, across platforms. followers is null for an account we have not read yet, and those come last. Each suggestion carries the profile id that /v1/creators/profile reads, and isAccount: true for the account of a brand or a page, false for a person who makes content, null before we judged it. An empty list costs $0. Suggestions carry no email; find contacts with /v1/contacts/find/fast or /v1/contacts/find/thorough.

GET /v1/creators/filter-options Filter values

The countries, niches, sub-niches and languages in the index, each with its profile count. Use these spellings in the search filters. Free: reference data costs nothing.

GET /v1/creators/avatars Avatar URLs

Profile picture URLs for up to 100 Instagram handles, as { handle: url }. The URL is our permanent mirror copy when we hold one, else the platform CDN URL, which expires. A handle we do not hold maps to null, and a call that finds no avatar costs $0.

POST /v1/creators/profiles/refresh-status Data freshness

How fresh our stored copy of each handle is, for up to 500 handles on one platform: known (we hold the account), last_scraped_at, candidate_status and refresh_pending (a crawl is queued). It reads state and never queues a crawl; fetch a fresh copy with the live operations.

POST /v1/creators/profiles/refresh Refresh a profile

Queues a fresh crawl of one stored profile. requested: true means this call queued it and bills $0.0005. requested: false is free: a crawl was already queued (state: pending), or the account is known not to exist (state: unavailable). Poll /v1/creators/profiles/refresh-status until refresh_pending is false, then read the fresh copy with /v1/creators/profile.

Intelligence

AI profile intelligence, custom enrichment, and brand and creator analytics.

POST /v1/intelligence/profile/visual-enrich Profile visual enrichment

Renders an Instagram account’s stored profile header and recent post grid as one JPEG and classifies it with a vision model: entity type, account role, niche and sub-niches, country, language, commercial intent, brand safety, and a visual read of the grid (format mix, aesthetic, palette, content pillars). The answer also carries the JPEG the model saw in data.image_base64 and the token counts. The result is returned, not stored. An account we do not store is a 404 and is not billed. The example image is truncated.

GET /v1/intelligence/profile Profile intelligence

Returns the AI intelligence bundle for a profile: entity type, niche classification, commercial intent, brand safety, demographics, location, language, and AI summaries (bio, content, audience, brand fit). Null fields are omitted. For emails use /v1/contacts/find. Returns the cached bundle when available, otherwise runs enrichment (scraping the profile first if needed), persists it, and returns it.

GET /v1/intelligence/brand/creators Brand collab creators

Creators who collaborated with a brand, ranked by distinct collab posts, with followers, mention_count and last_collab_at (null when no collab post was dated). Instagram edges are tags, co-authors, brand tags and paid partnerships; TikTok edges are caption @mentions.

Contacts

The contact finder: everything you know about a creator or brand goes in, and their emails (ranked by type and confidence, each with its evidence), social handles, websites and phones come out. fast answers within 2 seconds, thorough within about 30. POST /v1/contacts/find/{level} takes any subject; GET /v1/contacts/find-from-{instagram,tiktok,youtube}/{level} take one handle.

POST /v1/contacts/find/fast Find contacts (fast)

Everything you know about one creator or brand goes in: handles, profile URLs, emails, websites, phones or a name.

POST /v1/contacts/find/thorough Find contacts (thorough)

Everything you know about one creator or brand goes in: handles, profile URLs, emails, websites, phones or a name.

GET /v1/contacts/find-from-instagram/fast Find contacts from Instagram (fast)

The general find with one Instagram account as the subject.

GET /v1/contacts/find-from-instagram/thorough Find contacts from Instagram (thorough)

The general find with one Instagram account as the subject.

GET /v1/contacts/find-from-tiktok/fast Find contacts from TikTok (fast)

The general find with one TikTok account as the subject.

GET /v1/contacts/find-from-tiktok/thorough Find contacts from TikTok (thorough)

The general find with one TikTok account as the subject.

GET /v1/contacts/find-from-youtube/fast Find contacts from YouTube (fast)

The general find with one YouTube account as the subject.

GET /v1/contacts/find-from-youtube/thorough Find contacts from YouTube (thorough)

The general find with one YouTube account as the subject.

Batch

Async jobs over the whole catalog: submit an array of inputs against any batchable operation (batchable on its doc entry), poll progress, download the finished snapshot. One control plane — no per-operation batch endpoints. Rows are the exact single-call inputs (query params as a JSON object for GET operations, the request body for POST operations); each row bills at the normal single-call price as it executes, failed rows are free, and the control-plane calls themselves are free. Bright Data-compatible polling shape: trigger → progress → snapshot, 409 snapshot_not_ready until terminal, results retained 14 days.

POST /v1/batch/trigger Submit a batch job

Run an array of inputs against any batchable catalog operation. Each row is EXACTLY the input the single-call route takes — a JSON object of query params for GET operations, the request body for POST operations. Free to call; a flat pre-authorization (rows × single-call price) must fit your balance, and each row bills individually as it executes (failed rows are free). At most 10,000 rows, 4 MB body, and 5 active jobs per organization.

GET /v1/batch/progress/{snapshot_id} Poll a batch job

Free. status walks starting → running → ready (or failed / canceled); cost accrues as successful rows bill. A snapshot this key does not own is a 404.

GET /v1/batch/snapshot/{snapshot_id} Download batch results

Streams results in submission order once the job is terminal; 409 snapshot_not_ready while it runs. Every row carries its input alongside data (or error, with include_errors=true). Completed rows of failed/canceled jobs stay downloadable until expiry. For very large snapshots page with batch_size + part — each page is one bounded response.

DELETE /v1/batch/snapshot/{snapshot_id} Cancel a job / delete its results

On an active job: cancel — the dispatcher stops, in-flight rows finish (and bill), completed rows stay downloadable, status becomes canceled. On a terminal job: delete the stored results now (early retention); the job row remains for the usage ledger.

GET /v1/batch/snapshots List batch jobs

This account's jobs, newest first. Free.