Tiktok Scraper 2.0

Scrape TikTok profiles, user videos, keyword search results, video metrics, media links, and transcript text into clean Apify dataset rows.

This Actor is designed for monitoring TikTok accounts, collecting public video analytics, tracking keyword trends, and feeding structured TikTok data into dashboards, warehouses, or enrichment pipelines.

Features

• Scrape videos from TikTok usernames
• Scrape videos from TikTok keyword queries
• Collect profile metadata for user runs when enabled
• Extract video metrics including plays, likes, comments, shares, saves, duration, hashtags, and music title
• Extract subtitle transcript text when exposed by TikTok metadata
• Fall back to cleaned caption text when no transcript track is available
• Return compact pipeline-friendly rows by default
• Label row quality with metadata_quality and missing_fields
• Store continuation checkpoints between runs to skip already-seen videos
• Write run diagnostics to OUTPUT_SUMMARY

Workflows

Users

Use workflow: "users" to scrape one or more TikTok accounts.

{
  "workflow": "users",
  "users": ["tiktok", "khaby.lame"],
  "maxVideosPerUser": 10,
  "skipProfileScrape": false,
  "outputMode": "compact"
}

Keywords

Use workflow: "keywords" to scrape videos discovered from TikTok search and hashtag pages.

{
  "workflow": "keywords",
  "keywords": ["funny", "dog training"],
  "maxVideosPerKeyword": 10,
  "outputMode": "compact"
}

Input Reference

Field	Type	Default	Description
workflow	string	users	Required. Use users or keywords.
users	array[string]	-	TikTok usernames, @handles, or profile URLs. Required for users workflow. Max 100.
keywords	array[string]	-	Keyword queries. Required for keywords workflow. Max 100.
maxVideosPerUser	integer	10	Maximum videos to collect per user.
maxVideosPerKeyword	integer	10	Maximum videos to collect per keyword.
engine	string	auto	Primary browser engine: auto, playwright, or pydoll.
outputMode	string	compact	compact for smaller records, full for more technical fields.
maxDatasetItems	integer	5000	Maximum dataset rows pushed for the run. Use 0 for no additional cap.
skipProfileScrape	boolean	true	Faster users workflow. Set to false for richer account metadata.
includePlaybackUrl	boolean	true	Include direct playback URLs. These are often signed and short-lived.
includeTranscriptText	boolean	true	Include transcript or caption fallback text in rows.
includeMediaLinksMeta	boolean	true	Include expiry/signature metadata for media URLs.
enableApifyContinuation	boolean	true	Load checkpoint state from the named key-value store.
resetApifyContinuation	boolean	false	Ignore previous checkpoint state and replace it after the run.
continuationStoreName	string	tiktok-scraper-2-0-continuation	Named key-value store for checkpoints. Normalized to lowercase letters, digits, and hyphens.
continuationStateKey	string	CONTINUATION_STATE	Key used inside the continuation store.
minDelaySec	number	1.1	Minimum randomized browser delay.
maxDelaySec	number	2.4	Maximum randomized browser delay.
ytdlpTimeoutSec	integer	240	Timeout for metadata/transcript fallback calls.
ytdlpChunkSize	integer	40	Profile playlist items requested per yt-dlp chunk. Larger values can improve high-volume user runs.
ytdlpMaxRawScan	integer	200	Maximum raw playlist entries scanned per user. Use 0 for no additional scan cap.
ytdlpMaxVideosPerUser	integer	0	Optional yt-dlp-specific cap per user. Use 0 to follow maxVideosPerUser.
ytdlpDetailEnrichLimit	integer	10	Maximum per-video detail enrichments per user. Lower values are faster but may reduce metadata depth.
ytdlpTranscriptDetailEnrichLimit	integer	80	Maximum per-video detail enrichments used for transcript discovery.
ytdlpTranscriptSubtitleDownloadLimit	integer	8	Maximum subtitle-download transcript fallback attempts per user.
disableScraplingFallback	boolean	false	Disable the secondary fallback path.

Provide either users or keywords, not both.

Continuation

Continuation is enabled by default and helps avoid returning the same videos across repeated runs.

The Actor stores checkpoint state in the named key-value store configured by continuationStoreName, under the key configured by continuationStateKey.

For each user or keyword, continuation stores:

• latest known from_epoch
• latest continuation_token
• recent video_ids for duplicate suppression
• last scrape timestamp

To force a fresh run, set:

{
  "resetApifyContinuation": true
}

You can also provide a one-off lower bound directly in an item:

• tiktok|from=2026-01-01T00:00:00Z
• khaby.lame|from=1778336344
• funny|from=2026-01-01

An explicit from value in the input takes priority over stored continuation state.

Dataset Output

Each dataset item is a video row. Compact rows start with:

• video_id
• video_url
• caption_text
• transcript

Common fields include:

• create_time and create_time_epoch
• author_username and author_id
• hashtags
• play_count, like_count, comment_count, share_count, save_count
• duration_s
• music_title
• media_links
• transcript_detail
• has_transcript
• has_caption_text_fallback
• metadata_quality
• missing_fields
• account for user workflow rows
• search_keyword for keyword workflow rows

Transcript Fields

has_transcript is true when subtitle/caption metadata was exposed and parsed.

has_caption_text_fallback is true when no transcript track was available and the Actor used cleaned caption text instead.

transcript_detail.status may be:

• found
• fallback_description
• missing
• disabled

Metadata Quality

metadata_quality helps downstream systems decide whether a row is complete enough for analytics:

• full: core metadata is present
• partial: usable video row with some missing fields
• url_only: mostly just a discovered video URL

missing_fields lists important absent fields such as create_time_epoch, author_id, play_count, or duration_s.

Run Summary

Every run writes OUTPUT_SUMMARY to the default key-value store.

Useful fields include:

• status
• videos_scraped
• videos_with_transcript
• videos_with_caption_text_fallback
• videos_with_any_transcript_text
• metadata_quality_counts
• health
• continuation_loaded_entities
• continuation_loaded_id_entities
• continuation_updates
• dataset_items_pushed

The health object includes:

• total_rows
• sparse_row_count
• empty_entity_count
• seed_count
• empty_seed_count
• challenge_seed_count
• success_rate

Performance Tips

• Use outputMode: "compact" for most API and dataset workflows.
• Set includePlaybackUrl: false when you do not need direct signed playback URLs.
• Keep skipProfileScrape: true for faster user runs.
• Set skipProfileScrape: false only when you need richer account metadata.
• Use smaller maxVideosPerUser and maxVideosPerKeyword for frequent monitoring.
• Use continuation for repeated monitoring runs.
• Increase ytdlpTimeoutSec when transcript extraction is more important than speed.

For high-volume user monitoring where speed and lower compute usage matter more than transcript extraction, use compact rows and disable transcript/media enrichment:

{
  "workflow": "users",
  "users": ["tiktok", "netflix", "nba"],
  "maxVideosPerUser": 80,
  "maxDatasetItems": 1000,
  "outputMode": "compact",
  "includePlaybackUrl": false,
  "includeTranscriptText": false,
  "includeMediaLinksMeta": false,
  "skipProfileScrape": true,
  "ytdlpChunkSize": 80,
  "ytdlpMaxRawScan": 80,
  "ytdlpDetailEnrichLimit": 0,
  "ytdlpTranscriptDetailEnrichLimit": 0,
  "ytdlpTranscriptSubtitleDownloadLimit": 0
}

API Usage

Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

run_input = {
    "workflow": "users",
    "users": ["tiktok", "khaby.lame"],
    "maxVideosPerUser": 5,
    "skipProfileScrape": False,
    "outputMode": "compact",
    "includePlaybackUrl": False,
}

run = client.actor("thescrapelab/tiktok-scraper-2-0").call(run_input=run_input)

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["video_url"], item.get("play_count"), item.get("metadata_quality"))

HTTP

curl -sS -X POST \
  "https://api.apify.com/v2/acts/thescrapelab~tiktok-scraper-2-0/run-sync-get-dataset-items?format=json&clean=true" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "workflow": "keywords",
    "keywords": ["funny"],
    "maxVideosPerKeyword": 5,
    "outputMode": "compact",
    "includePlaybackUrl": false
  }'

Notes And Limitations

• TikTok changes page behavior often, and some runs can encounter challenge or login pages.
• Keyword scraping uses both search and hashtag discovery paths, but availability can vary by region and timing.
• Transcript tracks are only returned when exposed by TikTok metadata or fallback sources.
• Caption fallback is not a true spoken transcript.
• Direct media URLs can be signed and expire quickly.
• Public follower/following lists are not reliably available from TikTok public web pages.