TikTok Transcript Scraper: Captions & Speech to Text

Get the transcript of any public TikTok video: timestamped text from TikTok captions, or speech-to-text when a video has no captions. No TikTok login, no cookies.

Paste one or more TikTok video links (or a profile name) and get one row per video with clean text, segments with start and end times, AI-ready chunks and video metadata. The smallest run is one video link and takes a few seconds.

Table of contents

• What it does
• How it works
• Quick start
• Use cases
• Input
• Output
• Pricing
• Limits
• API, export and integrations
• Troubleshooting
• FAQ
• Related Actors
• Support

What it does

The Actor turns public TikTok videos into text. For each video it returns:

• the full transcript as plain text;
• segments with start and end times in seconds;
• chunks of about 1,000 characters with times, ready for AI tools;
• metadata: author, description, hashtags, date, duration, views, likes, comments, shares, saves and sound;
• a clear status and reason when a video has no text (photo post, library sound only, silence, deleted video and more).

Text comes from TikTok's own captions when the video has them. When it has none, the Actor listens to the audio and writes the text with Whisper speech-to-text (you can turn this off). It also can take the newest videos of a public profile and save SRT, VTT and TXT files.

Good fit

• Repurpose TikTok videos into blog posts, captions, newsletters or scripts.
• Feed video text to ChatGPT, Claude or a RAG pipeline for summaries and research.
• Track what creators or competitors say in their newest videos, on a schedule.

👉 Try it now: open the input form

How it works

1. The Actor opens the public page of each TikTok video, the same page anyone can see without an account. For a profile, it reads the list of its newest public videos.
2. If TikTok has captions for the video (written by the creator, made by TikTok's speech recognition, or TikTok's translation), the Actor uses them. Otherwise it downloads the audio for a moment and turns the speech into text. The video is not stored.
3. It saves one row per video in the dataset, with a status for every input link, and a short run summary (RUN_SUMMARY) in the key-value store.

👉 Try it now: open the input form

Quick start

1. Open the Input tab. The field TikTok video links already has one example video (box 1 in the picture).

2. Replace it with your own links, one per line. You can also add usernames in Profiles.

3. Keep Transcript options as they are (box 3): captions first, speech-to-text when there are none.

4. Click Start.

5. When the run ends, open the Output tab. The Overview table shows the status, where the text came from and the transcript (box 5).

   

Expected result: one row with status complete, source captions and about 100 words of English text with 12 timed segments.

👉 Try it now: open the input form

Use cases

Buyer job	Input path	Output entity	Key fields	Next integration
Turn a video into a blog post or script	videoUrls	One row per video	transcript, description, hashtags	Copy to a doc, or send to an LLM
Give an AI tool clean video text	videoUrls	Row with chunks	chunks (text, start, end), videoUrl	RAG / vector database, MCP agent
Make subtitles	videoUrls + saveSubtitleFiles	SRT / VTT / TXT files	files.srt, files.vtt	Video editor
Watch what creators say	profiles + skipProcessed	Rows for new videos only	transcript, createTimeISO, stats	Schedule, Slack or Notion summary
Research topics or trends	profiles or videoUrls	Rows with metadata	transcript, hashtags, stats.playCount	CSV / Excel

👉 Try it now: open the input form

Input

Input fields

Field	Plain name	Type	Required	Default / prefill	Allowed values	Effect	Example	Recommendation
videoUrls	TikTok video links	list of text	one of videoUrls / profiles	prefill: one example video	up to 1,000; video, short, photo and profile links, or video IDs	Which videos to transcribe	https://www.tiktok.com/@apifytech/video/7398101551744552225	Main input
profiles	Profiles	list of text	one of videoUrls / profiles	empty	up to 50 usernames or profile links	Takes the newest videos of each profile	@apifytech	For monitoring creators
maxVideosPerProfile	Videos per profile	number	no	10	1–500	How many newest videos per profile	20	Keep low for a first test
language	Caption language	text	no	empty (original language)	two-letter code	Uses TikTok captions in this language when they exist	es	Leave empty for the spoken language
fallbackToSpeechToText	Speech-to-text when no captions	yes/no	no	yes	yes / no	Transcribes audio of videos without captions	false	Turn off for the cheapest, fastest run
translateToEnglish	Translate to English	yes/no	no	no	yes / no	English captions, or Whisper translates the speech	true	When you need English for every video
maxSpeechMinutesPerVideo	Max speech minutes per video	number	no	10	1–60	Longer videos without captions get no text and no speech charge	5	Protects your budget
chunkCharacters	Chunk size	number	no	1000	0–10,000 (0 = no chunks)	Size of AI-ready chunks	2000	Match your AI tool's context
saveSubtitleFiles	Save subtitle files	yes/no	no	no	yes / no	Saves SRT, VTT and TXT files, links in files	true	For video editors
outputFormat	Output format	choice	no	nested	nested, flat	flat makes columns like authorUsername, playCount	flat	flat for CSV / Excel
skipProcessed	Skip videos from earlier runs	yes/no	no	no	yes / no	Skips videos already returned ($0.0001 each instead of a full row)	true	For scheduled runs
stateName	State name	text	no	tiktok-transcript-state	letters, digits, -	Store that remembers returned videos	brand-watch	One name per monitoring task
proxyConfiguration	Proxy	proxy	no	none	Apify Proxy or your own	Direct by default; switches to Apify Proxy by itself if TikTok refuses	—	Usually leave empty
deliveryConnector	Send summary to	connector	no	none	your Slack, Notion or other MCP connector	Posts a short summary after the run	—	Optional
deliveryTarget	Where to post	text	no	empty	Notion page link or Slack channel ID	Target page or channel	—	Optional

Give at least one link or profile; an empty input stops with a clear message. The same video given twice (for example with and without ?is_from_webapp=1) is processed once. Text that is not a TikTok link gets its own row with errorCode INVALID_URL. A profile link in videoUrls works like a profiles entry.

Interacting settings

When	Input A	Input B	Effective behavior	Example	Cost/result effect
A video has captions in several languages	language	translateToEnglish	translateToEnglish wins: English captions first, else Whisper translates	language: es, translateToEnglish: true	Whisper translation is charged like speech-to-text
A video has no captions	fallbackToSpeechToText	maxSpeechMinutesPerVideo	Transcribed only when on and the video is not longer than the limit	15-minute video, limit 10	No text (TOO_LONG), no speech charge
Scheduled monitoring	skipProcessed	stateName	Videos returned before under the same name are skipped	daily run of @brand	Skipped videos cost $0.0001 each (video-checked)
You set a maximum cost per run	Apify run option	fallbackToSpeechToText	Speech-to-text stops when the next unit does not fit; the last text may be cut (TRUNCATED)	max $0.05	You never pay more than you set

Input examples

One video

{
  "videoUrls": ["https://www.tiktok.com/@apifytech/video/7398101551744552225"]
}

Expected behavior: reads TikTok's captions of the video. Expected output: one row, status complete, source captions, transcript, 12 segments. Limit or cost note: the start fee ($0.005 at the default 2048 MB) and one video event, no speech charge.

Newest videos of creators, captions only

{
  "profiles": ["@apifytech"],
  "maxVideosPerProfile": 20,
  "fallbackToSpeechToText": false,
  "outputFormat": "flat"
}

Expected behavior: takes the 20 newest videos and uses only TikTok captions. Expected output: 20 rows; videos without captions have errorCode NO_CAPTIONS. Limit or cost note: only video events and a $0.0025 start fee, because the run uses 1 GB of memory; it is fast.

👉 Try it now: open the input form

Output

Output fields

Field	Meaning	Type	Can be empty?	Example / enum	Source / notes
videoUrl	Full link of the video	text	no	https://www.tiktok.com/@apifytech/video/7398101551744552225	Built from the author and video ID
inputUrl	The link as you gave it	text	no	https://vm.tiktok.com/…	Your input
videoId	TikTok video ID	text	when the link is invalid	7398101551744552225
status	Result of this row	text	no	complete, partial, not_found, error	partial = video found, no or cut text
source	Where the text came from	text	no	captions, whisper, none
captionType	Kind of text	text	yes	tiktok-creator, tiktok-asr, tiktok-translation, whisper-base
confidence	How sure the text is, 0–1	number	yes	0.85	Fixed for TikTok captions (creator 0.95, TikTok speech recognition 0.85, translation 0.75); measured for Whisper
language	Language of the text	text	yes	en
availableCaptionLanguages	Caption languages TikTok has	list	can be []	["en", "es"]
transcript	Full text	text	yes	Are you doing research in HR or …
segments	Timed lines	list of {start, end, text}	can be []	{"start": 0.02, "end": 1.42, "text": "Are you doing research in HR"}	Seconds
chunks	AI-ready pieces	list of {index, start, end, text}	can be []		Size from chunkCharacters
wordCount	Words in the transcript	number	0 when no text	96
noTranscriptReason	Why there is no or only part of the text	text	yes	The video uses only a sound from the TikTok library …	Plain English
errorCode	Short code of the reason	text	yes	see Troubleshooting
description, hashtags	Video caption text and hashtags	text, list	yes	#webscraping
createTimeISO, durationSeconds	Upload time (UTC) and length	text, number	yes	2024-07-31T…Z, 23
isPhotoPost	Photo post (slideshow)	yes/no	yes	false
author	username, nickname, id, verified, followerCount	group	yes	apifytech	Flat: authorUsername …
stats	playCount, diggCount (likes), commentCount, shareCount, collectCount (saves)	group	yes	783	Flat: playCount …
music	Sound title, author, isOriginal	group	yes	original sound - Apify	Flat: musicTitle …
files	Links to srt, vtt, txt files	group	yes		Only with saveSubtitleFiles
speechSecondsCharged	Seconds of audio charged for speech-to-text	number	0	60	Rounded up to 10 s
profile	Profile this video came from	text	yes	apifytech
fetchedAt	When the row was made (UTC)	text	yes	2026-11-01T07:40:07Z

Output entities and surfaces

Situation	Dataset/API/KVS/live view	Result shape	User action
Normal item	Dataset, views Overview and Transcripts with timestamps	status complete, text and metadata	Open, export or use transcript / segments
Video has no speech	Dataset	status partial, metadata, errorCode + noTranscriptReason	Read the reason; nothing to retry
Deleted or private video, bad link	Dataset	status not_found or error, errorCode	Check the link
Technical failure or retry	Dataset + RUN_SUMMARY in the key-value store	errorCode BLOCKED / SKIPPED_BLOCKED; summary reason SOURCE_ERROR	Run again later with the same input
Subtitle files	Key-value store, transcript-<videoId>.srt / .vtt / .txt	Files, links in files	Download or link

Output example

{
  "videoUrl": "https://www.tiktok.com/@apifytech/video/7698373675267001622",
  "videoId": "7698373675267001622",
  "status": "complete",
  "source": "captions",
  "captionType": "tiktok-asr",
  "confidence": 0.85,
  "language": "en",
  "transcript": "If I just makes it super easy to unlock that, because going to try and get a Google Maps API …",
  "segments": [
    {"start": 0.1, "end": 2.78, "text": "If I just makes it super easy to unlock that,"},
    {"start": 2.781, "end": 3.581, "text": "because"}
  ],
  "wordCount": 155,
  "noTranscriptReason": null,
  "errorCode": null,
  "hashtags": ["apify", "aiagents", "vibecoding", "mcp"],
  "createTimeISO": "2026-10-21T14:00:00Z",
  "durationSeconds": 57,
  "author": {"username": "apifytech", "nickname": "Apify", "verified": false, "followerCount": 3589},
  "stats": {"playCount": 783, "diggCount": 5, "commentCount": 0, "shareCount": 1, "collectCount": 4},
  "music": {"title": "original sound - Apify", "author": "Apify", "isOriginal": true},
  "files": null,
  "speechSecondsCharged": 0
}

This example is trimmed: it shows 2 of the segments and leaves out chunks and a few fields. null means the value does not apply (no error, no files asked). Text from TikTok's speech recognition or from Whisper can have small mistakes, like "If I" in place of "Apify" above.

Send results to Slack or Notion

Pick your connector in the optional deliveryConnector field. After the run saves its data, the Actor posts a short summary (up to 25 rows and a link to the full dataset) there. Put a Notion page link or a Slack channel ID in deliveryTarget: in Notion every run is added on top of that one page, so your workspace does not fill up with new pages. A delivery error never fails the run and costs nothing extra. Connect a service once in Apify Console → MCP connectors.

👉 Try it now: open the input form

Pricing

Pricing unit

You pay per event. Apify platform usage (compute, proxy, storage) is included: the developer pays it.

Event	When it is charged	FREE	BRONZE	SILVER	GOLD and higher
apify-actor-start	Once per run, one event per GB of run memory (minimum one)	$0.0025	$0.0025	$0.0025	$0.0025
video	A video was found and its row has metadata, with or without text	$0.0009	$0.00081	$0.00072	$0.00063
speech-10-seconds	Each started 10 seconds of audio sent to speech-to-text	$0.002	$0.0018	$0.0016	$0.0014
video-checked	TikTok was checked but there is no paid row: a video not found, a profile not found, private or without videos, a video skipped by skipProcessed after TikTok listed it	$0.0001	$0.0001	$0.0001	$0.0001

• The start fee grows with memory: $0.005 at the default 2048 MB, $0.0025 at 1024 MB (the Actor picks 1024 MB itself when fallbackToSpeechToText is off and translateToEnglish is off), $0.01 if you raise memory to 4096 MB. It covers starting the run, so a run that finds nothing is not free.
• Videos with TikTok captions cost the video event: $0.90 per 1,000 videos on the FREE plan, plus one start fee per run. Put many videos in one run to share it.
• Speech-to-text is charged for the audio it processed, also when the result is NO_SPEECH (music or noise). One minute of audio is 6 units: $0.012 on FREE.
• A found video without text (photo post, library sound only, silent video, NO_CAPTIONS, TOO_LONG) is charged as one video event, because you get its metadata and the reason. Speech-to-text is not started for these, so there is no speech charge.
• A deleted or private video, a profile that is not found, private or empty, and a profile video skipped by skipProcessed cost one video-checked event ($0.0001): TikTok still has to be asked. That is $0.10 per 1,000.
• Invalid links, blocked requests (SKIPPED_BLOCKED) and links skipped by skipProcessed before any request cost nothing.
• A run where nothing is found costs only the start fee ($0.005 at 2048 MB). A run where every link is invalid ends as Failed and also pays the start fee, because the run was started.
• If TikTok refuses 3 videos in a row on every route, or stops answering for over 45 seconds, the Actor stops sending requests: the remaining rows are SKIPPED_BLOCKED and cost nothing, so such a run costs about the start fee.
• If you set a maximum cost per run, the Actor stops before it. The last speech-to-text text may then be cut (TRUNCATED).

Cost examples

Measured runs on the FREE plan prices:

Scenario	Input scope	Expected item/event count	Base charge	Optional charge	Stop/cap behavior
One video with captions (prefill)	1 link, 2048 MB	2 apify-actor-start, 1 video	$0.0059	none	—
All links invalid	1 invalid link, 2048 MB	2 apify-actor-start	$0.005	none	run ends Failed
1,000 videos that are not found	1,000 links, 2048 MB	2 apify-actor-start, 1,000 video-checked	$0.105	none	took 140 s
1,000 videos, captions only	20 profiles × 50 videos, fallbackToSpeechToText: false, 1024 MB	1 apify-actor-start, 1,000 video	$0.9025	none	took 133 s
60 newest videos with speech-to-text	3 profiles × 20 videos, 2048 MB	2 apify-actor-start, 60 video, 154 speech-10-seconds	$0.059	$0.308	took 533 s
A 1-minute video without captions	1 link, 2048 MB	2 apify-actor-start, 1 video, 6 speech-10-seconds	$0.0059	$0.012	—

👉 Try it now: open the input form

Limits

• Only public videos and public profiles. Private accounts and private videos return PRIVATE_ACCOUNT or not_found. No login is used, so the Actor cannot see what only followers see.
• Up to 1,000 links and 50 profiles (up to 500 newest videos each) per run, but at most 1,000 videos in total per run (links and profile videos together). Profiles after that limit get a row with RUN_VIDEO_LIMIT; a profile not reached before the run timeout gets RUN_TIME_LIMIT. Split bigger jobs into several runs.
• Speech-to-text works on one video at a time: about 1 hour of audio takes about 22 minutes of run time. The default run timeout is 1 hour, which fits about 2.5 hours of audio. For more, raise the timeout or split the input.
• Speech-to-text uses the Whisper base model. It is good for clear speech; songs, strong accents and noise give weaker text (look at confidence).
• A video with only a sound from the TikTok library (a song, a popular sound) has no original speech: it gets NO_ORIGINAL_AUDIO and no text.
• Short links (vm.tiktok.com/…, tiktok.com/t/…) are supported; a short link that no longer leads to a video gets NOT_FOUND. We checked the expired case live; a working short link was checked only in tests.
• The output fields are our own. They are not the same as other TikTok transcript Actors, so a switch may need small changes in your code.
• Comments, video files and downloads are not part of the output. The video is not stored.
• TikTok can change its pages or refuse requests. The Actor retries through Apify Proxy; rows it still cannot get are marked SKIPPED_BLOCKED and cost nothing.

👉 Try it now: open the input form

API, export and integrations

• API: start a run with the same JSON as the input form and read the dataset items. See the Apify API docs.
• Export: download the dataset as JSON, CSV, Excel, XML or HTML. Use outputFormat: "flat" for tables.
• Scheduling: run on a schedule with profiles and skipProcessed: true to get only new videos each time.
• Webhooks / no-code: use Apify integrations (Make, Zapier, n8n, webhooks) to send each new row on; the next system reads transcript and videoUrl.
• Subtitles: with saveSubtitleFiles: true the SRT, VTT and TXT links are in files.

Use with AI agents (MCP)

• MCP server URL with only this Actor: https://mcp.apify.com?tools=cliqtomedia/tiktok-transcript.

• Smallest useful input:

  {"videoUrls": ["https://www.tiktok.com/@apifytech/video/7398101551744552225"]}
  

• Key output fields: transcript (full text), segments (timed lines), chunks (pieces with times for long context), status and noTranscriptReason (why a video has no text), videoUrl.

• One video with captions takes a few seconds and costs $0.0059 on FREE ($0.005 start fee at 2048 MB + $0.0009 per video). Videos without captions take about 3 seconds of run time per 10 seconds of audio and add $0.002 per 10 seconds.

• Limits: up to 1,000 videos per run, public videos only, speech-to-text up to maxSpeechMinutesPerVideo (10 min by default) per video.

👉 Try it now: open the input form

Troubleshooting

A video has no transcript

Symptom: status is partial and transcript is empty. Likely reason: see errorCode and noTranscriptReason:

errorCode	Meaning
NO_CAPTIONS	TikTok has no captions and fallbackToSpeechToText is off
NO_ORIGINAL_AUDIO	Only a TikTok library sound, no original speech
NO_SPEECH	Silent video, or only music or noise
NO_AUDIO	The video file has no sound
PHOTO_POST	Photo slideshow, no spoken audio
TOO_LONG	Longer than maxSpeechMinutesPerVideo
CHARGE_LIMIT	Your maximum cost per run left no room for speech-to-text
RUN_TIME_LIMIT	The run reached its timeout first
ABORTED	You aborted the run first
TRUNCATED	Speech-to-text stopped at your maximum cost per run; the text is cut
DOWNLOAD_ERROR	The video file did not download (larger than 300 MB, empty or a download error)
LOGIN_REQUIRED	TikTok gives this video's file only after login
NETWORK_ERROR, BLOCKED	TikTok did not answer or refused the request for the video file

Check: the reason text. Action: turn on fallbackToSpeechToText, raise maxSpeechMinutesPerVideo, the run timeout or the maximum cost, as the reason says. For download, login or network errors, run the video again later; these rows cost one video event and no speech charge.

A link returns not_found or error

Symptom: status is not_found (NOT_FOUND, PRIVATE_ACCOUNT, NO_VIDEOS) or error (INVALID_URL, RUN_VIDEO_LIMIT). Likely reason: the video was deleted, the account is private, the text is not a TikTok link, or the run already took 1,000 videos. Check: open the link in a private browser window. Action: fix or remove the link, or move extra profiles to another run. not_found rows cost one video-checked ($0.0001); error rows cost nothing. If every link is invalid, the run ends as Failed and pays only the start fee.

The run ends with SOURCE_ERROR

Symptom: some rows have SKIPPED_BLOCKED and RUN_SUMMARY shows reason SOURCE_ERROR. Likely reason: TikTok refused requests even through Apify Proxy, refused 3 videos in a row, or did not answer for over 45 seconds (NETWORK_ERROR, then SKIPPED_BLOCKED). Check: the rows that are complete are saved and charged as normal. Action: run again later with the same input; use skipProcessed to avoid the videos you already have.

👉 Try it now: open the input form

FAQ

Do I need a TikTok account or cookies?

No. The Actor reads only public pages, without login, cookies or your own proxy.

Why is speech-to-text charged when the result is NO_SPEECH?

Speech-to-text has to process the whole audio to learn that it is only music or noise. That work is charged per 10 seconds. Videos that clearly have no speech (library sound only, silent, photo posts) are not sent to speech-to-text.

Is the transcript complete and correct?

It is as good as TikTok's captions or the Whisper base model. TikTok's own speech recognition and Whisper can mishear names and words. confidence and captionType show where the text came from.

Can I get the text in another language?

Set language to use TikTok's captions in that language when TikTok has them, or translateToEnglish to get English for every video.

👉 Try it now: open the input form

Related Actors

• YouTube Transcript Scraper: Subtitles & Timed Text — the same kind of timed text for YouTube videos.
• Audio & Podcast Transcriber (Speech to Text) — speech to text for audio and video files from other sources.
• Social Video Downloader: TikTok, Instagram, X - No Login — when you need the TikTok video file itself.

👉 Try it now: open the input form

Support

For a bug, question or feature request, open the Actor's Issues tab on Apify. Include the run ID, the input links that failed and the errorCode. Never send tokens, cookies or private data.

👉 Try it now: open the input form