TikTok Hashtag Scraper

Give it up to 10 TikTok hashtags and it returns the videos posted under them, one row per video. Rows use the same field names as clockworks/tiktok-hashtag-scraper (authorMeta, videoMeta, searchHashtag, ...). It is not a like-for-like replacement: read the next section before you rely on it.

Read this first: what works today

Live hashtag scraping is often blocked. Since 2026-09-02 TikTok has walled its hashtag pages (/tag/<name>) for datacenter IP addresses. A live crawl needs Tokfluence's ISP proxies to get through, and we cannot promise that tier is available at any given time. When every proxy we try is blocked, that hashtag comes back empty and the run's status message says WALLED and names the hashtag. If the last attempt failed some other way (a captcha, a timeout), the message names that error instead. The run still succeeds and you are only charged for rows you receive.

Stored hashtag feeds are thin. In database mode (and in auto mode for a hashtag we crawled in the last 7 days) you get only the videos Tokfluence has already seen under that hashtag. Tokfluence keeps a hashtag's videos only from live crawls made through its scrape API, which is new, so expect few videos or none for most hashtags, not the full feed. A hashtag's stored feed grows only when it is crawled live.

So for now expect fewer videos per hashtag than TikTok shows, and some runs with none. When a run comes back short, the status message says what we know about why.

What it returns

One video row per video, the same row shape as Clockworks' hashtag scraper. Filled when the stored video has them:

• id, text, textLanguage, createTime, createTimeISO, webVideoUrl, isAd
• diggCount, playCount, commentCount, shareCount, collectCount, repostCount
• hashtags (the caption's tags), mentions, detailedMentions and effectStickers (id and name)
• musicMeta (null when the video has no sound) and videoMeta (size, duration, format, cover, subtitle links)
• authorMeta.id, authorMeta.name, authorMeta.profileUrl, and the author's fans, following, heart, video and friends counts as they were when we captured the video
• searchHashtag.name: the hashtag this video was found under, without the # and in lower case
• input: the hashtag as you typed it

A video found under two of your hashtags appears once per hashtag, each row with its own searchHashtag, so it counts as two results.

Fields Tokfluence adds that Clockworks does not have sit under a tokfluence object:

• tokfluence.scraped_at: when Tokfluence last captured this video, so you can judge how fresh the counters are
• tokfluence.is_branded and tokfluence.ad_disclosure: Tokfluence's own reading of the caption for brand deals and ad labels. This is our classifier, not TikTok's sponsored flag, which is why isSponsored is null.

Inputs

• hashtags (required): up to 10 per run, with or without the #. Case does not matter; #Crochet and crochet are the same hashtag and are sent once.
• resultsPerPage: the most videos to return per hashtag, 100 by default, at most 1000.
• maxResults: the most rows for the whole run.

Fresh or fast: the Mode setting

Under Advanced, mode trades freshness for speed:

• auto (default): serves what Tokfluence already stores for a hashtag when it saw a video under that hashtag in the last 7 days, and crawls TikTok live for the rest. Most runs want this. A hashtag that goes live and is blocked comes back WALLED, even if we hold older videos for it.
• database: fastest. Never scrapes, so you get only the videos we have already seen under the hashtag (see above), newest sighting first, and their counters are as old as our last visit (check tokfluence.scraped_at). This is the way to get what we hold for a hashtag that came back WALLED.
• live: always crawls TikTok now. Freshest when it gets through; slower, and subject to the wall above.

A live crawl reads the top of the hashtag page, a fixed number of scrolls, not the whole feed, so a large resultsPerPage can come back short even when nothing is blocked. Every video a live crawl finds is stored, so a database run afterwards returns them without crawling again. Live crawls run on Tokfluence's shared scraping workers, one hashtag after another, and wait behind other live requests, so a live run can take a while.

If a live crawl outlasts the run's timeout, the run fails with the request id. The crawl keeps going and is stored when it finishes, so a database run shortly afterwards returns those videos.

Fields that can be null

This actor never fills a gap itself: when we do not have a field it is null, not 0, false or an empty string. Always null today:

• searchHashtag.views: we do not store a hashtag's total view count.
• isMuted, isPinned, isSponsored, isSlideshow, locationCreated: not captured. hasTikTokShopProduct: not filled by this actor.
• mediaUrls, slideshowImageLinks, commentsDatasetUrl: Clockworks fills these with files it copies; we do not copy files.
• url, error, errorCode, invalidUrls, submittedVideoUrl, fromProfileSection, searchQuery, searchMusic: they describe error rows or other Clockworks inputs. A hashtag we could not scrape is explained in the status message instead of in an error row.
• In authorMeta: nickName, avatar, signature, bioLink, commerceUserInfo, ttSeller and region, because a hashtag row carries the video, not the author's profile. Also verified, privateAccount, digg, isUnderAge18, roomId, createTime, originalAvatarUrl and followDatasetUrl.
• In musicMeta: coverMediumUrl, originalCoverMediumUrl, playUrl, musicAlbum.
• In videoMeta: transcriptionLink, and downloadLink inside subtitleLinks. The TikTok subtitle URL is in tiktokLink and is signed, so it stops working some time after the scrape.
• In hashtags: title and cover.
• In detailedMentions: nickName and postUrl.
• In effectStickers: stickerStats.

Often null:

• hashtags[].id and detailedMentions[].id: filled only when TikTok returned them with the video.
• videoMeta.coverUrl falls back to TikTok's own cover, which expires, when we have not stored a copy yet. videoMeta.originalCoverUrl and videoMeta.downloadAddr are often empty in the video as we stored it.
• locationMeta: only when the video is tagged with a place.

Zero or short results

When a run returns nothing, or fewer rows than maxResults, it still succeeds, but it logs a warning and sets the run's status message to the likely cause:

• WALLED: TikTok blocked every proxy we tried on that hashtag's page. Switch Mode to database for what we already hold, or try again later.
• a hashtag that does not exist, or is not a valid hashtag;
• another error code for a hashtag (for example CAPTCHA_REQUIRED), which is usually worth retrying later;
• stored feeds being thin (database mode, or a hashtag auto served from storage);
• a live crawl reading only the top of the hashtag page;
• resultsPerPage times the number of hashtags being below maxResults.

The actor uses one Tokfluence account for every run. If that account is out of API credits, or Tokfluence's scrape service is down, the run fails with a message saying so. You are charged only for rows already in the dataset.

What Tokfluence keeps from a run

Tokfluence logs every request the actor makes: the hashtags, the mode, where each row was served from, and a reference made of a one-way hash of your Apify user id plus the run id. Every hashtag you give, and every TikTok handle and hashtag in the returned rows, is recorded as a candidate for Tokfluence's creator discovery. Every video a live crawl finds is stored in Tokfluence like any other scrape, and later runs, anyone's, can be served that stored copy.

Memory

The actor defaults to 256 MB and allows up to 512 MB. It makes one API call, maps the rows and writes them to the dataset in batches of 100.

Pricing

PLACEHOLDER (Daniel): pay per result, one result event per dataset row. Price to be set in Console at publish.

If your plan's remaining budget covers fewer rows than you asked for, the actor collects up to that ceiling, says so in the log, and stops cleanly rather than returning a silently short list.

Notes

Data comes from public TikTok pages, scraped and stored by Tokfluence. You are the data controller for anything you export; follow GDPR and any local rules that apply to you. More at Tokfluence.