Douyin AI Transcript Scraper 🎙️

Scrape and transcribe Douyin (抖音) videos into accurate, speaker-labeled text. Paste one or more video URLs, and get full text transcripts — with automatic language detection, speaker identification, and smart formatting.

What does Douyin AI Transcript Scraper do?

This Actor takes one or more Douyin video URLs and returns a full text transcript of the spoken audio in each video. It automatically detects the spoken language (Chinese, English, or any of 30+ supported languages), identifies different speakers, and formats numbers, dates, and punctuation.

Only Douyin video URLs are accepted — non-Douyin links (YouTube, TikTok, etc.) are rejected before any processing starts. Each video produces one dataset row with the transcript and key video metadata such as the caption, creator name, and engagement stats.

You can access results through the Apify API, schedule runs on a timer, or connect them to tools like Google Sheets, Slack, Zapier, and Make.

Why scrape Douyin video transcripts?

• Content repurposing — Turn Douyin video content into blog posts, articles, or social media text.
• Market research — Analyze spoken content from Chinese social media at scale.
• Accessibility — Generate subtitles and captions for Douyin videos.
• Competitive intelligence — Monitor what competitors or influencers say in their videos.
• Translation prep — Get the original spoken text before you translate it to other languages.
• SEO and content indexing — Make video content searchable by extracting the spoken words.
• Trend analysis — Track popular topics and keywords across Douyin creators.

How to scrape Douyin video transcripts

1. Go to the Actor's Input tab on the Apify Console.
2. Paste one or more Douyin video URLs into the Douyin Video URLs field.
3. (Optional) Turn Speaker Diarization on or off. It is on by default.
4. (Optional) Turn Smart Formatting on or off. It is on by default.
5. Click Start and wait for the run to finish.
6. Open the Dataset tab to view or download your transcripts.

Accepted URL formats

The Actor accepts only Douyin video URLs. Non-Douyin links are rejected immediately — no network call is made and no charge is incurred.

Format	Example	Where you find it
Standard video page	https://www.douyin.com/video/7045159024525905183	Browser address bar on douyin.com
Standard video page (no www)	https://douyin.com/video/7045159024525905183	Same page, some browsers drop the www
Share link	https://v.douyin.com/iYRfMBam/	Douyin app → Share → Copy Link
Legacy share link	https://www.iesdouyin.com/share/video/7045159024525905183	Older Douyin app versions

Not accepted:

• URLs from other sites (YouTube, TikTok, Bilibili, etc.)
• Douyin user profile pages (douyin.com/user/...)
• Douyin search or hashtag pages
• Douyin image/photo posts (detected and skipped with a clear error message)
• Plain text that is not a URL

Query parameters and trailing slashes in the URL are fine — the Actor strips them automatically.

Input parameters

Field	Type	Description	Default
urls	string[]	Douyin video URLs or share links to transcribe (see accepted formats above)	(required)
maxAudioMinutes	integer	Maximum audio length in minutes (1–600)	600
diarize	boolean	Label different speakers in the transcript	true
smartFormat	boolean	Format dates, times, numbers, and punctuation	true

Example input

{
    "urls": [
        "https://www.douyin.com/video/7045159024525905183",
        "https://v.douyin.com/iYRfMBam/"
    ],
    "diarize": true,
    "smartFormat": true
}

Output — Douyin video transcript data

Each video produces one dataset row with the transcript and video metadata. You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.

For single-URL runs, the transcript is also saved as a plain-text file in the key-value store.

Example output

{
    "videoUrl": "https://www.douyin.com/video/7045159024525905183",
    "desc": "Video caption text",
    "authorName": "Creator Name",
    "createTime": "2024-01-15T10:30:00+00:00",
    "durationSeconds": 45.2,
    "language": "zh",
    "transcript": "[Speaker 0] 大家好，欢迎来到我的频道...",
    "transcriptLength": 1250,
    "transcriptStatus": "ok",
    "speakerCount": 2
}

Douyin transcript data fields

Field	Description
videoUrl	The Douyin video URL
desc	Video caption and description
authorName	Creator's display name
createTime	Upload date and time (ISO 8601)
cover	Thumbnail image URL
durationSeconds	Video length in seconds
transcript	Full transcript text with speaker labels
transcriptLength	Character count of the transcript
transcriptStatus	Result status: ok, empty, or an error code
language	Auto-detected spoken language code (e.g. zh, en)
speakerCount	Number of speakers detected
model	AI model used for transcription (e.g. nova-3)
diarize	Whether speaker diarization was enabled

How much does it cost to transcribe Douyin videos?

The Actor uses Pay Per Event pricing with two charge events:

1. Video processed ($0.005) — A flat fee charged once per video when processing starts.
2. Second of video transcribed ($0.00025/second) — Charged per second of audio, with a 30-second minimum. Only charged when a transcript is delivered.

Example costs:

Video length	Cost
30 seconds	$0.0125
2 minutes	$0.035
10 minutes	$0.155

Invalid URLs are rejected before any charge. Videos that cannot be fetched incur only the flat per-video fee.

Tips for better Douyin transcripts

• Batch processing — Paste multiple URLs at once. Each video gets its own dataset row.
• Language detection is automatic — The Actor detects the spoken language. You do not need to specify it.
• Share links work — Both v.douyin.com/xxx/ share links and full douyin.com/video/ID URLs are accepted. See the Accepted URL formats section for the full list.
• Invalid URLs are caught early — Non-Douyin URLs are rejected before any network call. No charge is incurred for rejected URLs.
• Empty transcripts — If the result is empty, the audio likely has no spoken words (for example, music, sound effects, or silence). The Actor retries automatically before it reports an empty result.
• Long videos — You can transcribe videos up to 10 hours long. Use the Max Audio Minutes setting to set a cap.

FAQ

Is it legal to scrape Douyin? This Actor accesses publicly available data. Always review Douyin's Terms of Service for your use case.

What languages does the transcription support? The Actor supports 30+ languages with automatic detection. Chinese (Mandarin) and English produce the best results.

Why is my transcript empty? The video audio likely has no spoken words — for example, pure music, singing, or silence. Speech-to-text does not transcribe sung lyrics.

Can I schedule automatic runs? Yes. Use the Apify Console scheduler to run the Actor on a timer with the same input.

Need help or have feedback? Open an issue on the Issues tab or contact the developer.