Douyin Transcripts: speech-to-text, 抖音 video to text avatar

Douyin Transcripts: speech-to-text, 抖音 video to text

Deprecated

Pricing

from $3.00 / 1,000 audio minute transcribed (whisper base)s

Go to Apify Store
Douyin Transcripts: speech-to-text, 抖音 video to text

Douyin Transcripts: speech-to-text, 抖音 video to text

Deprecated

Transcribe Douyin (抖音) videos to text or timestamped segments, by video URL, share link or ID, with each video's caption, hashtags, date, author and likes. Speech-to-text by Whisper inside the run; Mandarin in Simplified Chinese, and about 100 other languages.

Pricing

from $3.00 / 1,000 audio minute transcribed (whisper base)s

Rating

0.0

(0)

Developer

Spyridon Pikoulas

Spyridon Pikoulas

Maintained by Community

Actor stats

0

Bookmarked

3

Total users

2

Monthly active users

2 days ago

Last modified

Features

Limited permissions

Share

Retired on 2026-11-15. Douyin's operator has won suits against those who took data from it, and its user agreement prohibits crawlers, so this Actor no longer reads Douyin. A run reads nothing, charges nothing and ends with this notice.

Get the transcript of Douyin (抖音) videos as clean text, or as timestamped segments, from their URLs, the share links the app copies, or their IDs. Mandarin comes out in Simplified Chinese, and about 100 other languages are recognised too.

The speech-to-text runs inside the actor: OpenAI's Whisper model, open source, on Apify's servers. Nothing is sent to an outside AI service, and no Douyin account or cookies are needed.

Every result also carries what Douyin shows about the video: its caption, hashtags, date, duration, likes, comments, shares and saves, the author's name and Douyin ID, the sound used, and Douyin's label where the author declared the video AI-generated.

It costs $0.008 per minute of audio on Apify's paid plans ($0.016 on the free plan) with Whisper small, the default, and $0.003 ($0.006) with the faster Whisper base, counted in started minutes: a 45-second video is $0.008. Videos without speech are free.

What people use it for

  • Market research in China: what sellers, brands and creators say about products, as text you can search, translate and analyse.
  • AI and LLM pipelines: summaries, translation, RAG over Douyin videos, training data.
  • Content and trend research: the scripts behind videos that sell or go viral.
  • Subtitles and repurposing: timestamped lines for subtitles, quotes and translations.

How to transcribe a Douyin video

  1. Click Try for free; Apify's free plan covers a first run.
  2. Paste video URLs, share links or IDs under Douyin videos. The whole text the app's Share → Copy link gives works as it is.
  3. Click Start, then download the results as JSON, CSV, Excel or HTML, or read them through the API.

Max minutes per video transcribes, and charges, only each video's first minutes. Speech-to-text model picks between Whisper small, more accurate, and Whisper base, about three times faster and cheaper but with more wrong characters in Mandarin. Language fixes the language spoken when you know it; otherwise each video's is detected. Turn on Include timestamps for segments, each line with its start and end.

{
"urls": [
"https://www.douyin.com/video/7685708860878769454",
"https://www.douyin.com/jingxuan?modal_id=7702256834182253862",
"7693093964693490980"
],
"model": "small",
"includeTimestamps": true
}

Output

{
"input": "https://www.douyin.com/video/7685708860878769454",
"id": "7685708860878769454",
"url": "https://www.douyin.com/video/7685708860878769454",
"description": "2026新手版!AI设计必备整合包重磅更新! … #AI视频 #AI生成视频",
"hashtags": ["AI视频", "AI生成视频"],
"createdAt": "2026-09-15T10:35:45Z",
"durationSeconds": 43,
"likes": 341,
"comments": 5,
"shares": 17,
"saves": 114,
"authorName": "Ai工作流辣妹",
"authorDouyinId": "25677521181",
"authorUrl": "https://www.douyin.com/user/…",
"musicTitle": "@Ai工作流辣妹创作的原声一Ai工作流辣妹",
"label": "作者声明:内容由 AI 生成",
"language": "zh",
"languageProbability": 1,
"text": "兄弟们大得来了 这是一期老破晓也能无限生成电商设计的工作流 首先第一个必须给到H3多参生视频工作流 想做产品宣传 又没有模特的回专业团队 …",
"transcribedSeconds": 43.3,
"isComplete": true,
"billedMinutes": 1,
"segments": [{"start": 0, "end": 1.16, "text": "兄弟们大得来了"}]
}

A video stopped short by Max minutes per video, the run's spending limit or its timeout comes with what was transcribed, isComplete: false, its transcribedSeconds and a stoppedBecause, and is charged only for those minutes. A video without speech (music only), a photo post, or a video that's deleted or private comes with what's known of it and a noTranscriptReason, and costs nothing.

Use it as an API

Apify's API starts a run and returns its dataset as JSON; its Python and JavaScript clients, and its Make, Zapier and n8n integrations, work the same way.

curl -X POST "https://api.apify.com/v2/acts/pikoulas~douyin-transcripts/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
-H "Content-Type: application/json" \
-d '{"urls": ["https://www.douyin.com/video/7685708860878769454"]}'

AI agents can call it as a tool through Apify's MCP server: https://mcp.apify.com?tools=pikoulas/douyin-transcripts.

How much does it cost to transcribe Douyin videos?

With Whisper small you pay $0.008 per started minute of audio on Apify's paid plans, and $0.016 on the free plan: most Douyin videos are under a minute or two, so $8–16 per 1,000 videos ($16–32 on the free plan). Whisper base is $0.003 per started minute ($0.006 on the free plan). billedMinutes shows each video's. Videos without speech are free. If Douyin refuses Apify's servers, the actor reads its pages through Apify's residential proxy, charged at $0.0012 per 100 KB: about 40 KB a video, $0.0005. The video files themselves come straight from Douyin's servers, at no charge. There are no monthly fees.

The run stops at the spending limit you set, mid-video if need be, and charges only what it transcribed.

Good to know

  • Whisper writes what it hears: names, brands, slang and fast speech can come out as the wrong characters, Whisper base more often than small. Speakers aren't told apart, and captions burned into the video aren't read, only speech.
  • Chinese is written in Simplified characters, converted from Whisper's output, which mixes in Traditional ones.
  • A run transcribes about 3 minutes of audio a minute with Whisper small at the default 4 GB of memory, and about 8 with Whisper base. More memory is faster, at the same price per minute.
  • Videos are read one by one, by URL or ID; a creator's whole profile, search results and comments aren't read.
  • Deleted, private and friends-only videos can't be read; the result says so.
  • A video given twice in one run is transcribed once.

FAQ

It reads only what Douyin shows publicly, without logging in. Videos are their creators' copyrighted work: a transcript for your own research, analysis or accessibility is often allowed, depending on your country's copyright law, but republishing it may need the rights holder's permission, and personal data in it is protected by laws such as the GDPR, the CCPA and China's PIPL. Douyin's user agreement prohibits crawlers, and your use of it is your own responsibility. If you're unsure, ask a lawyer; Apify's Is web scraping legal? is a good start.

Not affiliated with Douyin or ByteDance; their names are trademarks of their owners.

Something missing or not working?

Open an issue on the Issues tab with the input you used, or ask there for a field you need.

If it does what you need, a review on the Reviews tab helps others find it.

Transcripts from other sites?

YouTube, TikTok, Instagram Reels, podcasts (Apple Podcasts and RSS) and any other video or audio URL each have an Actor of their own.