YouTube Subtitles Scraper — Transcripts & Captions at Scale

Extract YouTube captions and transcript text from watch URLs or video IDs. Get tracks, timed cues, availability flags, and a joined text preview into an Apify Dataset — built for NLP, search indexes, and compliance archives, not one-off “copy transcript” clicks.

Run YouTube Subtitles Scraper on Apify · Built by ScrapeMamba · Public captions only · Not a media downloader

---

Why this Actor?

Phone “Show transcript” works for one video. Catalogs, warehouses, and weekly jobs need batch automation: many IDs, exportable JSON/CSV, and schedules. This Actor is the suite’s YouTube transcript extractor — same videoId keys as Video, Search, and Related scrapers.

• Batch videoIds in one run
• Preferred caption lang (default en)
• Optional expandCues for one Dataset row per cue
• text preview truncated to 2000 characters; full cues stay on the subtitles row
• Clear signals when tracks exist but timed cues are withheld (cuesError)

---

Key features

• Bulk YouTube transcript / caption extraction
• Tracks metadata + timed cues when available
• Optional flat cue rows for warehouses that prefer timestamp tables
• Per-video error isolation
• Fits after Search / Video / Related shortlisting

---

What you can scrape

Subtitles row (type: "subtitles")

Group	Fields
Identity	videoId, lang
Availability	isAvailable, source, trackCount, cueCount
Content	tracks, cues, text (preview ≤ 2000 chars)
Diagnostics	cuesError / related flags, raw, error

Cue row (type: "cue", when expandCues: true)

videoId, id, start, duration, text, language, languageCode

---

Use cases

• NLP / topic models — Speech text beside video metadata
• Content research — Quote mining and theme extraction
• Compliance archives — Periodic public caption snapshots
• Search indexes — Index spoken content for internal discovery
• Accessibility audits — Check which catalog IDs expose captions

---

Example output

{
  "type": "subtitles",
  "videoId": "dQw4w9WgXcQ",
  "lang": "en",
  "isAvailable": true,
  "source": "youtube",
  "trackCount": 1,
  "cueCount": 180,
  "text": "We're no strangers to love…",
  "tracks": [{ "languageCode": "en", "name": "English" }],
  "cues": [
    { "start": 0.0, "duration": 3.5, "text": "We're no strangers to love" }
  ]
}

With expandCues: true, each cue also becomes its own Dataset row (type: "cue").

---

How to use

1. Open YouTube Subtitles Scraper
2. Pass one known-captioned public video; keep expandCues: false for the first run
3. Scale the videoIds list once the schema looks right

Input example

{
  "videoIds": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"],
  "lang": "en",
  "expandCues": false
}

API (Node.js)

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('scrapemamba/youtube-subtitles-scraper').call({
  videoIds: ['dQw4w9WgXcQ'],
  lang: 'en',
  expandCues: false,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();

Local run

cd youtube-subtitles-scraper-js
cp .env.example .env
npm install
node src/main.js

---

Input parameters

Parameter	Required	Default	Description
videoIds	Yes	—	Watch URLs, Shorts, youtu.be, embeds, or 11-character IDs. Free users limited to 200 videos per run; paying users uncapped.
lang	No	"en"	Preferred caption language code
expandCues	No	false	If true, also push one Dataset row per cue (multiplies billable rows)

---

Output

Default runs produce one type: "subtitles" row per video. expandCues: true adds type: "cue" rows. API failures become type: "error" while other videos continue.

The Actor cannot invent captions when none are exposed. Auto-generated and human tracks both count when YouTube provides them.

---

Why choose this Actor

• Public captions only — Clear availability flags; cuesError when tracks exist without timed cues
• Flexible storage — Keep cues as JSON on the parent row, or expand to flat rows
• Suite join keys — Same videoId as Video / Search / Related
• Batch transcript automation — Built for catalogs, not phone copy-paste

---

Pricing

$1.99 per 1,000 results on Apify. Cue expansion multiplies rows. Free users: first 200 videos per run. See the Store page.

---

FAQ

How do I get YouTube transcripts in bulk?

Pass many URLs/IDs, set lang, keep expandCues: false unless you need cue rows, run on Apify, export the Dataset.

What if captions are off?

If no track (including auto-generated) exists, the Actor will not invent speech — expect empty availability or an error.

What does cuesError mean?

Tracks were found, but timed cues were not returned. Inspect raw; try another language or retry. Different from “captions fully off.”

Auto vs human captions?

Both are returned when available. Quality varies; the Actor does not guarantee a human track.

Do Shorts work?

Yes, if you pass Shorts URLs/IDs and captions exist.

How many videos can I process in one run?

Free users are limited to the first 200 videos per run; paying users are uncapped.

Should I store cues as JSON or rows?

JSON on the subtitles row is usually cheaper. Use expandCues when your warehouse prefers flat timestamp tables.

---

Related Actors

Actor	Use it for
YouTube Video Scraper	Metadata beside captions
YouTube Search Scraper	Find videos to caption
YouTube Related Videos Scraper	Expand a neighborhood before selective captioning
YouTube Channel Videos Scraper	Shortlist uploads to caption

Use only on public caption data per YouTube’s terms, copyright/privacy law, and your policies. Do not bypass disabled captions or access private videos.