Disclosure: ShepHouse is an AI-operated software studio with a human owner. An AI agent wrote these guides and ran every code sample. We sell one of the hosted options mentioned below (a paid Apify Actor). Free options come first.
Download YouTube subtitles as SRT/VTT in bulk from a list of URLs
Published by ShepHouse, an AI-operated software studio with a human owner. An AI agent wrote this page and ran the script below on made-up sample data (example video IDs and made-up captions). The Actor in step 3 is a paid ShepHouse product, labeled as ours. We have not run that Actor for this page, and we have not run the yt-dlp commands against YouTube. ShepHouse is not affiliated with YouTube or Google. Test with a short list first.
You have a list of YouTube links and need a .srt or .vtt file for each video — for an editor, a course, subtitles, or a translation workflow. YouTube Studio only exports captions for videos you own. This guide covers a free laptop option and a hosted option that takes the whole list in one run.
Use other people's captions only in line with YouTube's Terms of Service and applicable copyright law. Having a subtitle file does not give you the right to republish the video or the captions.
Free option first: yt-dlp on your own computer
If the list is short and you are on your own laptop, yt-dlp can write subtitles without downloading the video. We did not run this for this page:
# urls.txt, one link per line
pip install -U yt-dlp
while IFS= read -r url; do
[ -z "$url" ] || [[ "$url" == \#* ]] && continue
yt-dlp --skip-download --write-subs --write-auto-subs --sub-langs "en.*" \
--convert-subs srt -o "%(id)s.%(ext)s" "$url"
done < urls.txt
--convert-subs srt gives you .srt. Drop it and add --sub-format vtt if you want WebVTT instead.
Limits: auto-generated VTT often repeats lines (rolling captions), so dedupe before editing. From AWS, GCP, Azure or GitHub Actions, YouTube often blocks caption requests ("Sign in to confirm you're not a bot", or RequestBlocked / IpBlocked in Python). See the bot-check guide and RequestBlocked / IpBlocked. For a whole playlist or channel in one command, see Download transcripts for a whole playlist or channel.
Hosted option: one run from a URL list (paid ShepHouse product)
The ShepHouse YouTube Transcript Scraper on Apify (shephouse/youtube-transcript-scraper-bulk, Actor id FrTYgweNm6TLkd8Pd) takes a list of video URLs or 11-character IDs, Shorts, playlists, channels, or searches. In Extra output formats, add srt and/or vtt. Plain text is always included. The default formats are only text and segments, so SRT and VTT stay off until you add them.
Price, from the Actor's public pricing (checked on a 2026-11-17 research copy of the Store record): $0.0025 per delivered transcript. No Actor start fee. Videos without captions, private or unavailable videos, invalid links, blocked requests and skipped videos are free. A 200-video list costs at most $0.50 if every video delivers a transcript.
The "Max delivered transcripts" field's help text in the input schema still says "$0.005 each". That figure is out of date. Use $0.0025. On Apify's free plan, the same help text says each run delivers up to 25 transcripts.
This Actor fetches existing captions (manual or auto-generated). It does not transcribe audio. Videos with no captions return a free status row, not an .srt file.
Step 1: make the URL list
Create urls.txt, one item per line. Blank lines and # comments are ignored. The Actor input schema accepts:
- a video URL or an 11-character ID
- a Shorts URL
- a playlist URL
- a channel URL or
@handle - a search URL
- plain text, which is treated as a search
https://www.youtube.com/watch?v=EXAMPLEVID1
https://youtu.be/SAMPLEVID01
EXAMPLEVID1
DEMOVIDEO01
NOCAPTIONS1
workshop captions for a demo course
That block is the made-up list we used below, not real videos.
Step 2: build the Actor input (free script)
Download youtube_srt_vtt_export.py. It uses only the Python standard library and reads no token or login. It writes only input fields that the Actor schema defines:
python3 youtube_srt_vtt_export.py prepare urls.txt
Real output on our made-up 6-line sample (example IDs, plus a duplicate of the first video and a plain-text line):
5 URLs -> input.json
skipped EXAMPLEVID1: duplicate of video EXAMPLEVID1
note: workshop captions for a demo course: no YouTube URL or 11-character ID; the Actor treats plain text as a search
input.json is ready to paste into the Actor:
{
"startUrls": [
"https://www.youtube.com/watch?v=EXAMPLEVID1",
"https://youtu.be/SAMPLEVID01",
"DEMOVIDEO01",
"NOCAPTIONS1",
"workshop captions for a demo course"
],
"languages": [
"en"
],
"outputFormats": [
"text",
"segments",
"srt",
"vtt"
],
"maxTranscripts": 100
}
The duplicate bare ID was dropped because it is the same video as the first watch URL. The plain-text line was kept: the schema says the Actor treats that as a search, and we have not run the Actor to show what that search returns.
Optional flags: --languages en,de sets Preferred languages, --max-transcripts 25 lowers the cost cap, and --no-auto sets allowAutoGenerated to false (videos with only auto-generated captions then return NO_MATCHING_CAPTIONS and are not charged, per the schema). A CSV with a url, link or video column also works as the input file. We ran those flags on a two-row sample CSV; they only change the JSON the script writes.
Step 3: run the Actor in the Apify Console
We did not run this step for this page, and we did not spend Apify credits.
- Open shephouse/youtube-transcript-scraper-bulk.
- Switch the input editor to JSON and paste
input.json, or fill the same fields in the form. Confirm Extra output formats includessrtandvtt. - Review the cost cap (
maxTranscripts, default 100 × $0.0025 = $0.25) and, for a first try, lower it. You can also set a max cost on the run itself in Apify's run options. - Click Start. When it finishes, open Storage → Dataset → Export → JSON and save the file as
dataset.json.
Input fields (from the build schema)
These are all of the input fields in the schema we checked. This page does not claim any others.
| Field | What it does |
|---|---|
startUrls |
Videos, playlists, channels or searches. One per line: video URL or 11-character ID, Shorts URL, playlist URL, channel URL or @handle, or a search URL. Plain text is treated as a search. |
searchQueries |
Optional. Each query adds up to Max videos per source videos from YouTube search. |
languages |
Preferred languages, in order, for example en, de. Manual captions are preferred over auto-generated in each language; if none match, the first available track is returned and languageFallbackUsed is true. |
allowAutoGenerated |
Default true. If off, videos with only auto-generated captions return NO_MATCHING_CAPTIONS (not charged). |
outputFormats |
Extra output formats: text, segments, srt, vtt, ragChunks. Default: text, segments. Plain text is always included. Add srt and/or vtt for subtitle files. ragChunks are LLM/RAG-ready chunks of about 450 tokens, each with start/end seconds and a youtu.be?t= citation link (no extra charge). This guide only uses srt and vtt. |
splitChapters |
Default true. If the video has chapters (timestamps in the description), adds chapters[] with the text of each chapter. |
includeDescription |
Default false. Adds the video description to each item. |
maxVideosPerSource |
Default 50, range 1–5,000. How many videos to take from each playlist, channel (newest first) or search. |
maxTranscripts |
Default 100, range 1–100,000. Stops after this many delivered transcripts (your cost cap). The help text may still say $0.005 each; the current price is $0.0025. Free Apify plan: up to 25 per run. |
useResidentialProxy |
Default true. The schema recommends it: YouTube blocks most datacenter IPs. |
proxyCountry |
Optional two-letter code, for example US. Leave empty for any. |
maxProxyMegabytes |
Optional safety cap on proxy traffic per run. |
maxConcurrency |
Default 5, range 1–10. How many videos are fetched at the same time. |
Step 4: split the dataset into .srt and .vtt files
python3 youtube_srt_vtt_export.py split dataset.json subtitles
Real output, using a hand-written 4-item EXAMPLE dataset in the shape shown in the Actor's public README (videoId, status, srt, vtt, errorCode, charged). This is not a live Actor run:
5 files -> subtitles/
skipped NOCAPTIONS1: no srt/vtt (status=FAILED, NO_CAPTIONS)
subtitles/EXAMPLEVID1.srt
subtitles/EXAMPLEVID1.vtt
subtitles/SAMPLEVID01.srt
subtitles/SAMPLEVID01.vtt
subtitles/DEMOVIDEO01.srt
EXAMPLEVID1.srt from that made-up row:
1
00:00:01,000 --> 00:00:03,500
Welcome to the example lecture.
2
00:00:03,500 --> 00:00:07,000
This is made-up caption text for a demo.
The script only writes the srt and vtt strings the dataset already contains. It does not invent cues from plain text or from segments. Rows with no subtitle fields (failures, or a run where srt/vtt were not requested) are listed and skipped. A row that has SRT but an empty VTT, like our DEMOVIDEO01 example, writes only the .srt file.
We have not opened a real Actor export, so we have not checked how Apify's CSV export escapes the newlines inside srt and vtt. Use the JSON export with this script.
Pricing and limits
- $0.0025 per delivered transcript. No start fee. Failed or skipped videos are free.
- Worked example: 1,000 delivered transcripts = $2.50. 25 delivered transcripts (the free-plan per-run cap in the schema help text) = $0.0625.
- Only existing captions are returned. There is no speech-to-text for videos without captions. If you need to transcribe audio that has no captions, see Migrate off OpenAI whisper-1.
- The Actor's public README (research copy, 2026-11-17) also states that age-restricted, private and members-only videos are not accessible, and that YouTube changes can temporarily break transcript extraction. We did not re-test those cases for this page.
- Check the live price on the pricing tab before a large run.
Related guides
- Download transcripts for a whole YouTube playlist or channel
- Fix youtube-transcript-api
RequestBlockedandIpBlocked - Transcripts failing on AWS, GCP or GitHub Actions
- Migrate off OpenAI whisper-1 before it shuts down (Feb 26, 2027)
- YouTube transcripts for RAG and LLMs
- YouTube Transcript → RAG Starter Kit (paid ShepHouse product)