YouTube Transcript Scraper · Chinese-first (中文影片逐字稿)

YouTube Transcript Scraper extracts YouTube video transcripts and subtitles in any language — with first-class Traditional and Simplified Chinese support. Paste video, channel or playlist URLs, click Start, and download transcripts as JSON, CSV or Excel, or as SRT/VTT subtitle files. 以中文為核心,也支援其他語言。

Other transcript scrapers fail on most Chinese videos, because most Chinese videos simply have no captions to scrape. We measured it: only ~25–37% of Chinese YouTube videos have any caption track, and auto-captions are almost never generated for Chinese. This Actor scrapes the captions when they exist — and, if you add your own free Groq key, transcribes the audio with AI when they don't.

What can this YouTube transcript scraper do?

• 📝 YouTube transcript API without the hassle — captions from single videos, whole channels or playlists (1–3 seconds per video)
• 🇹🇼 Chinese-first language picking — for human-made captions: Traditional → Simplified → generic Chinese → English by default, configurable for any language. 人工字幕:繁中 → 簡中 → 通用中文 → 英文
• 🗣️ Real transcripts, not machine translations — for videos with only auto-captions you get the language actually spoken in the video by default; YouTube's auto-translated tracks are skipped (switch Auto-captions in original language off to get them). 只有自動字幕時,預設給影片原本講的語言,不給機器翻譯(可關閉)
• 🎬 SRT / WebVTT subtitle files — optional, free, with timestamps. 可選 SRT/VTT 字幕檔,免費附加
• 🤖 Optional AI fallback — no captions? Transcribe the audio with Whisper using your own free Groq key
• ✅ Quality-gated — too short, error pages or wrong-language tracks are reported as failures, never silently returned or charged. 品質過濾,不交垃圾資料
• ⚙️ Apify platform features — API, scheduling, webhooks, integrations (Make, Zapier, n8n, Google Sheets), MCP for AI agents

How to scrape YouTube transcripts

1. Click Try for free and open the Input tab.
2. Paste YouTube video, channel or playlist URLs (or 11-character video IDs) into Video URLs.
3. Adjust Language preference if you want a language other than Chinese first (e.g. en, ja, es).
4. Optional: pick Subtitle files (srt / vtt); tick AI fallback and paste your Groq key for videos without captions.
5. Click Start, then export the dataset (JSON, CSV, Excel, HTML) from the Output tab or fetch it through the API.

How much does it cost?

Pay per event — you pay only for successful results, platform usage is included.

Event	Price	When
Caption transcript	$0.005 per video	The video has a caption track (manual or auto)
AI transcript	$0.03 per video	No captions; AI fallback on with your Groq key, transcription succeeded
Actor start	$0.00005	Tiny platform start fee

Examples: a 50-video channel with captions ≈ $0.25. SRT/VTT files are free. Failed or skipped videos cost nothing. 字幕檔免費;失敗不收費。

Input

Field	Description
videoUrls	Video, channel or playlist URLs (11-character IDs also accepted) 影片/頻道/播放清單網址
preferredLanguages	Caption language priority; Traditional Chinese first by default 語言優先序
autoCaptionOriginalLanguage	Default true: auto-caption-only videos come in the spoken language. false: pick by preferredLanguages, which may return YouTube's machine translation 自動字幕給原文(預設開)
aiFallback	Transcribe with AI when there are no captions (needs your Groq key) 沒字幕時用 AI
subtitleFormats	Also return srt / vtt subtitle files (free) 字幕檔輸出
groqApiKey	Your Groq key, only needed for AI fallback Groq 金鑰
maxItems	Cap when expanding a channel or playlist 頻道展開上限

Example input:

{
  "videoUrls": ["https://www.youtube.com/watch?v=pDxi9MljStw", "https://www.youtube.com/watch?v=aircAruvnKk"],
  "preferredLanguages": ["zh-Hant", "zh-TW", "zh-Hans", "zh", "en"],
  "subtitleFormats": ["srt"],
  "maxItems": 50
}

Output example

{
  "videoId": "pDxi9MljStw",
  "url": "https://www.youtube.com/watch?v=pDxi9MljStw",
  "title": "EP103 很怕別人過得比我爽",
  "channel": "蔡康永",
  "duration": 3821,
  "uploadDate": "20260714",
  "viewCount": 284119,
  "language": "zh-Hans",
  "source": "caption:manual",
  "charCount": 26985,
  "transcript": "簡單的來講就是不放心 不放心 我不放心別人是不是在過比我爽的很多的生活…"
}

source tells you where the text came from: caption:manual (human captions), caption:auto (auto-captions) or ai:groq-whisper-turbo (AI transcription).

Subtitle files (SRT / VTT) — free add-on

Pick srt and/or vtt in Subtitle files and each result also carries timestamped subtitles:

{
  "segments": [{"start": 1.0, "end": 3.5, "text": "大家好 歡迎收看本集節目"}],
  "srt": "1\n00:00:01,000 --> 00:00:03,500\n大家好 歡迎收看本集節目\n\n",
  "vtt": "WEBVTT\n\n00:00:01.000 --> 00:00:03.500\n大家好 歡迎收看本集節目\n\n"
}

Timestamps come from the YouTube caption file or from the AI transcription segments. If a source has no timing information, the transcript is still delivered with a subtitleNote. Leave the field empty and the output stays plain. 時間軸來自字幕檔或 AI 分段;留空=輸出不變。

Enabling AI fallback · AI 補位怎麼開

1. Get a free key at console.groq.com (Groq offers a free tier; check their site for current limits). 到 console.groq.com 申請免費金鑰
2. Tick aiFallback and paste the key into groqApiKey. 勾選 aiFallback,把金鑰貼進 groqApiKey

Your key is used for that run only and never stored. 金鑰只在該次執行中使用,不會被儲存。YouTube sometimes blocks audio downloads from cloud servers; when that happens the video is reported as failed and not charged.

Use cases · 誰在用

• Content research & competitor analysis — transcribe a whole channel for topic and keyword analysis 一次抓整個頻道做主題分析
• AI pipelines — clean text for RAG, summarization, Q&A and LLM agents AI 工作流餵料
• Subtitles & translation — get the original text and SRT, then translate or proofread 字幕與翻譯
• Accessibility — add text to Chinese content that has no captions 無障礙
• Language learning — Chinese/English transcripts with timestamps

Related transcript Actors

All four share the same quality gate (no charge for junk) and a similar JSON output — pick the one built for your source:

Actor	Best for
YouTube Transcript Scraper · Chinese-first (this one)	YouTube videos, channels, playlists — captions in any language, Traditional/Simplified Chinese first
Any Video to Transcript (AI)	Mixed lists: Instagram, X/Twitter, Vimeo, Dailymotion, Reddit, Bilibili and 1,000+ sites, built-in AI (no key)
TikTok Transcript Scraper	TikTok videos — reads TikTok's own captions ($0.002), AI when there are none
Facebook Video Transcript	Facebook videos, Reels, Watch, fb.watch

FAQ

Can I get a YouTube transcript in languages other than Chinese?

Yes. By default, videos with only auto-captions are returned in the language actually spoken (English, Japanese, Spanish…), whatever your preference list says. For human-made captions, put language codes in Language preference (e.g. ["en"], ["ja", "en"]); the first track that exists is used. Chinese is just the default.

Why don't I get a Chinese transcript of an English video?

YouTube offers machine-translated auto-caption tracks in 150+ languages. They are translations of speech recognition, not transcripts, so by default this Actor skips them and returns the original spoken language. Human-made translated captions are still used when they match your preference list. 英文影片若只有自動字幕,預設給英文原文(機器翻譯軌不算逐字稿)。

How do I get YouTube's machine-translated Chinese (or another language) instead?

Set Auto-captions in original language (autoCaptionOriginalLanguage) to false. Auto-captions are then picked by your Language preference list, so an English video can come back as YouTube's machine-translated Chinese. Check the language field to see which track was used. 想要 YouTube 機器翻譯的中文:把「自動字幕給原文」關掉(false),就會照語言優先序挑。

Does it work for whole YouTube channels and playlists?

Yes. Paste a channel or playlist URL and set Max videos (maxItems) to control how many videos are expanded.

What if a video has no captions?

Without AI fallback, the video is reported as failed in the run log and not charged. With AI fallback on and your Groq key, the audio is transcribed with Whisper and charged as an AI transcript.

Do I need a YouTube API key or Google account?

No. Only the optional AI fallback needs a (free) Groq key.

Traditional vs Simplified Chinese — which one do I get?

For human-made captions, the first available track in your preference list. By default Traditional Chinese (zh-Hant, zh-TW, zh-HK) is tried first, then Simplified (zh-Hans, zh-CN), then generic zh, then English. The language field tells you which track was used. For Chinese videos with only auto-captions, you get YouTube's original speech-recognition track by default (which may be Simplified); set autoCaptionOriginalLanguage to false to pick by your preference list instead. 只有自動字幕的中文影片,預設給 YouTube 原始辨識軌(可能是簡中);要照偏好挑繁中請把該選項關掉。

Can I download SRT or VTT subtitle files?

Yes — choose srt and/or vtt in Subtitle files. Free, no extra charge.

Why did some videos fail?

Typical reasons: no caption track (and AI fallback off), private or age-restricted video, or YouTube blocking the request. Failures are classified (proxy / rate-limit / blocked / network / other) in the run log and the RUN_REPORT record, and never charged. Keep Apify Proxy enabled — YouTube blocks datacenter IPs.

Can I use it from code, Make, Zapier, n8n or an AI agent?

Yes — use the Apify API (run-sync-get-dataset-items), the Python/JavaScript clients, the built-in integrations, or the Apify MCP server so AI agents can call it as a tool.

Is it legal to scrape YouTube transcripts?

This Actor only reads publicly available captions and audio. You are responsible for how you use the data — respect copyright, YouTube's terms and privacy laws such as GDPR. If unsure, consult a lawyer.

Something doesn't work — how do I get help?

Open an issue in the Issues tab with the run link and the video URL. We read every issue. 有問題請到 Issues 分頁留言。

Technical notes · 技術細節

Captions are fetched in the structured json3 format first (most reliable), falling back to vtt/srt with timestamp markup stripped. Retries depend on the status code (404 gives up immediately, 429 waits longer, 403/5xx back off). Errors are grouped into proxy / rate-limit / blocked / network / other in the run report. 字幕優先取 json3 結構化格式,退回 vtt/srt 並自動清掉時間軸標記;失敗依狀態碼分流重試。