Every YouTube video that has captions is quietly carrying a free transcript you can download — the exact timing, the exact dialogue, in a format you can edit, translate, or drop into any video player. You just need the right tool to unpack it. Whether you want the words from a lecture, a tutorial, or an interview, this guide shows you how to download subtitles from YouTube as clean SRT files in under a minute — no installs, no watermarks, no account walls.
We'll cover the fastest method first (paste the URL, done), then the browser-only method YouTube itself offers, then what to do with the file afterward: fixing auto-caption typos, converting languages, timing corrections for re-editing, and the legal etiquette around other people's captions.
Table of contents
What you get when you download YouTube captions
The fast way: paste the video URL into a subtitle downloader
The built-in way: Google's Takeout / transcript viewer (and why neither is a downloader)
Auto-generated vs. human captions — which tracks exist and what to expect
Downloading captions in other languages (and translating them properly)
What to do with the SRT afterward (edit, sync, convert, burn)
Legality and etiquette of downloading other people's captions
Troubleshooting: no captions found, garbled text, missing languages
FAQ
Before doing anything, it helps to know what the "subtitle" you're downloading actually is.
YouTube doesn't ship subtitle files with videos the way DVDs did. Captions live in YouTube's own database, attached to a video by either:
The uploader (human captions, in one or more languages, often the cleanest source),
YouTube's automatic speech recognition (auto-generated captions, available in roughly a dozen major languages — English, Spanish, French, German, Hindi, Arabic, etc.), or
The uploader's community or uploaded translations (some channels provide multi-language tracks).
A downloader's job is to convert YouTube's caption storage into a standard file — most usefully SRT, sometimes VTT or plain text. The output looks identical to any other SRT: numbered cues, timestamps, dialogue lines.
Why people download from YouTube, in rough order of frequency:
Transcripts for notes and study — lectures, conference talks, tutorials.
Subtitles for offline playback — you have the video and want matching captions in VLC or on a TV.
Accessibility — someone who can't enable YouTube's player captions (rare, but it happens in embedded/locked environments).
Translation workflows — grab the English track, translate it, re-time the result over the original video.
Content repurposing — pulling quotes, writing summaries, cleaning auto-captions into a script draft.
Research and analysis — dialogue, keywords, datasets.
Note what's not on that list: making money ready-made by scraping spoken content word-for-word. That distinction has real weight in section 7.
[IMAGE: A simple diagram of YouTube's caption system: video → tracks (human + auto) → downloader → SRT file]
This method solves the problem before YouTube's interface can slow you down, and it's the one we recommend first for any URL-shaped task.
The best tool for this specific job is the YouTube subtitle downloader at JustSubtitles. What makes it the go-to: it takes the raw video URL as the input — not a clip ID, not some renamed file — and it runs entirely in your browser. Here's the full workflow:
Copy the URL of the YouTube video. Works with youtube.com/watch?v=..., youtu.be/..., and even Shorts URLs. If the video is embedded somewhere or you're on mobile, still copy the share link — same thing.
Open the downloader — JustSubtitles' YouTube tool. No login, no email gate.
Paste the URL into the input box and press the button. The tool looks up every caption track attached to that video.
Pick your format and language. You'll get a list of available tracks — typically English (auto-generated) plus every language the uploader or community added. Choose per-track, and choose the format: SRT for players and editing, plain text if you just want the words.
Download. A clean .srt file lands in your Downloads folder, ready for VLC, an editor, a translation tool, or the recycle bin of your choice.
That's the whole loop: copy → paste → pick → download. Total time, 15–40 seconds once you know the drill. Because it's browser-based, the exact same flow works on Windows, Mac, Linux, Android, iPhone — no desktop app, no "free trial," no watermark on your file.
[IMAGE: Four-panel strip: the YouTube video with URL bar highlighted → pasted into the downloader box → a list of available languages → the SRT file saved]
It downloads the captions that exist. If a video has no captions at all (rare on talky content, common on music videos), there's nothing to download — section 8 covers what to do then.
It works on auto-generated and human caption tracks alike. Auto tracks come with the usual "speaking clearly" quirks covered in section 4.
Shorts work too, as long as they have captions.
Unlisted videos work (anything with a URL), but private videos you don't have permission for won't resolve — which is correct behavior, not a bug.
Age-restricted or region-locked videos sometimes fail; that's YouTube's gatekeeping, harmless and obvious.
No video is downloaded. This is purely a caption retrieval — small request, small file. Nothing touches the video stream itself.
Two features often confused with actual downloading; both worth knowing, neither giving you a file worth using.
On most captioned videos, expand the description or the three-dot menu and choose Show transcript. YouTube displays a scrollable transcript with timestamps, next to the video. You can copy the text out by selecting it.
But: the transcript is a viewer, not a file. What you get when you copy it is a wall of text — timestamps at inconsistent intervals — that no player can read, no editor imports cleanly, and re-timing loses the cue boundaries. It's fine for skimming what a talk says. It's not what you want to feed to anything.
[IMAGE: The "Show transcript" panel open on a YouTube video, next to a downloaded SRT file in Notepad showing proper cue format]
Similarly, YouTube's caption menu (the CC button and gear icon) switches tracks on and off and changes appearance — but there's no "export" action in there. You cannot download a caption file from the player UI no matter how deep you dig.
(There was a Takeout-era route — Google Takeout could export your own uploaded videos' caption tracks — but that covered your own channel, which is a case you can handle by copy-pasting from Studio anyway; and the API route that tool used has shrunk over the years.)
So: for actually having a file, a URL-based downloader is the way. The transcript viewer is for reading on-screen; the downloader is for keeping.
When a video's track list shows multiple English options, they're usually not duplicates. Understanding what each is saves you editing pain afterward.
Created by YouTube's speech recognition. Ever-present on talky videos, available in a dozen-plus major languages. Their virtues and sins:
Fast and always there — no uploader effort required.
Punctuation is either absent or approximate. Sentence boundaries get fuzzed; proper nouns get garbled ("Kubernetes" can arrive as "coober netties"; brand names get mangled).
No speaker labels. Whoever talks, it all reads as one continuous voice.
*Timing is usually good,* because it comes from the same audio analysis — auto tracks stay in sync even when their words drift.
Occasionally outright wrong in ways that matter — numbers, IDs, URLs.
Grammar rule of thumb: auto captions are a rough draft you're probably keeping anyway, because they're the only track that exists.
Made by a person (or the channel's team), or by a captioning service. Sins are rare and virtues are real:
clean punctuation, speaker labels, correct names;
real sentence-case English, not the wall-of-lowercase auto style;
sometimes SDH (includes [music], [door slams] cues).
When a human track exists for your language, prefer it. If both exist and your downloader lists both, the human one is usually the one without the "(auto-generated)" tag in the name.
For big channels, you'll see non-English tracks that are actually translations by the uploader. These have human quality — but sometimes the timings are for the human English track, so if the English was re-timed later and translations weren't refreshed, you can get a mismatched track. Not common, but if a foreign-language track seems weirdly timed against the video, this is why.
YouTube's caption system often has tracks that nobody on Earth would call "the subtitles" — that multilingual list in the downloader is one of the most quietly useful features of the whole tool.
The default behavior: if the video is in Hindi and you want Hindi captions, pick the Hindi track, which is almost always either human (uploader) or auto-generated in Hindi. Both exist. Download normally.
Several paths, in increasing order of quality:
Use an existing translated track if YouTube lists one (community or uploader-uploaded). Instant and usually good.
Use YouTube's own translate feature in the downloader, where supported — some tools request YouTube's machine translation on the fly. Quality is what you'd expect (serviceable, not polished) but the file is there in seconds.
Download the source-language track, then translate the text yourself. Get the SRT, strip the timings aside for a moment (a subtitles-to-text tool makes this a one-click step), run the plain text through a translator, then re-attach the timing file later. This two-step route gives you the most editing control, and it's what people do when the other two aren't available for their language pair.
One caution on path 2 and 3: machine-translated subtitle files read wrong after the round-trip — tense drift, dropped subjects, no punctuation. Usable for personal comprehension; not what you'd publish as "the English subtitles" of someone else's video.
[IMAGE: The language dropdown open in the downloader showing a mix of "(auto-generated)" and human tracks in multiple languages]
The file's real journey starts after the download. The free subtitle tool suite covers most of what routinely needs doing:
Fix auto-caption text. Open the SRT in any text editor and correct names, punctuation, or word errors. SRT is just text — nothing mystical about editing it.
Convert for the platform you're feeding. VTT for HTML5 players and web projects, plain text for transcripts and translation inputs, PDF for scripts and reading material. Each is a one-click conversion.
Fix timing. Auto-caption timing is usually right; if you're re-timing for a re-edited version of the video (different intro cut, etc.), shift the whole file's offset with an online tool in one step.
Merge or split cues. Famous auto-caption sins: 7-word lines, or one cue spanning 20 seconds. A subtitle merge/split tool re-cuts them to human reading pace.
Load it into your player. Put it next to your local copy of the video with the same base name (Interview.Name.en.srt beside Interview.Name.mp4), and VLC/Plex/Kodi/smart TVs will pick it up with no ceremony.
A habit worth building: keep your caption files in SRT anywhere you can (UTF-8, always), because it's the format every tool above takes without an argument. The zoo of formats and conversions around it is a separate read — our guide on converting subtitle formats covers it for when the file needs to go somewhere picky.
Straight talk, not a lecture.
What you're downloading is someone else's text. Captions — especially human-written ones — are copyrighted creative work in their own right, separate from the video they accompany. YouTube's transcription of the audio is likewise a derivative of someone else's speech.
Personal use is the quiet reality. Downloading a video transcript to study it, extract facts, or caption your own private copy of a lecture is what practically everyone does with these tools and is not the kind of thing that consumes anyone's attention. It's for your reading.
Where it stops being fine:
Re-uploading someone else's captions as your own on YouTube or anywhere else. Credit or don't — it's their writing.
Using auto-captions (or any track) as wholesale "content" — publishing the word-for-word transcript of a podcast or video as an article. Even for free content, that's strip-mining someone else's labor.
Scraping captions in bulk or building a dataset around someone else's caption library without checking the pipeline's rules.
The adjacent-but-worse things — using the same URL-paste pattern against content you clearly shouldn't have (paywalled, private, age-restricted you don't qualify for). A subtitle tool isn't a loophole for access control.
The rule of thumb: captions are for the accessibility and comprehension they were made for — not a free content harvest. Use them like a reader, not like a republisher, and you'll never need to know where the line actually sits in law.
"No captions available for this video." The video genuinely has none — the uploader turned off captions, YouTube didn't generate them (short/musical/audio-only content sometimes skips ASR), or the video type blocks them. Nothing to download; if you need the text anyway, the options are live transcription services or plugins — outside this guide's scope.
The track is there but empty or one giant cue. Some uploader tracks are broken or misformatted by YouTube's ingest. Try the auto-generated track for the same language instead — that pair (human empty, auto fine) happens daily.
Garbled text (é-style characters). Classic encoding mismatch after a conversion or old-tool edit. Re-download (fresh files are UTF-8) or resave the file as UTF-8 in your text editor. Full detail lives in our subtitle-format guide.
My language isn't in the track list. The video simply doesn't have it. If you need it as a viewer, YouTube's player-side auto-translate can machine-translate captions on the fly — but remember the player is a viewer, not a downloader. To keep a translated file, use the two-step translate-and-retime flow from section 5.
The SRT works in VLC but not on my TV / Roku / Instagram. The universal fix in every subtitle problem: convert to SRT (you're usually already there), and on TV/Roku match the file name to the video's base name. For social platforms, captions must be burned into the video — no sidecar files. Our guide to adding subtitles across devices covers the whole compatibility table.
"The captions don't match the video anymore." If the uploader re-edited or re-uploaded the video with new timings, captions ride alongside or break. A partial-fit subtitle is a shift fix away (constant offset) or hopeless if whole sections moved — re-download on the off chance the track was refreshed, then shift.
Q: How do I download subtitles from a YouTube video? Paste the video's URL into a downloader like the YouTube subtitle downloader at JustSubtitles, pick the language and format, download the SRT. Fifteen to forty seconds, browser-based, works on any device.
Q: Can I download auto-generated (AI) captions too? Yes. Auto tracks show in the same list as human captions — usually labeled "(auto-generated)". Same download, same SRT output, with the usual AI-transcription quirks in the text.
Q: Does this work on YouTube Shorts? Yes, when Shorts have captions — same paste-URL flow, same output. Shorts without captions (music compilations, etc.) have nothing to download.
Q: Can I download captions from unlisted or private videos? Unlisted: yes, if you have the link. Private: no — and that's correct behavior, not a bug; private means the owner's permission matters.
Q: Which format should I choose, SRT or plain text? SRT for anything that plays video or feeds a caption editor. Plain text if you only want the words for notes, translation, or reading. SRT converts to plain text in one click whenever you want, so default to SRT and extract text later.
Q: Are YouTube auto-captions accurate enough to use as subtitles? For offline viewing: yes, comfortably. For publication or permanent captions: expect to proofread. Auto tracks carry no punctuation, mangled proper nouns, and no speaker labels; a five-minute cleanup turns them into publishable files.
Q: Is downloading YouTube subtitles legal? Downloading for personal reading, study, or offline accessibility is what these tools exist for and nobody bothers with. Publishing, repurposing, or redistributing someone else's captions or transcripts is where it crosses into territory the law and YouTube's rules both care about.
Q: The video has captions on YouTube but my language isn't listed — why? Tracks exist only for languages someone uploaded or YouTube auto-detects. YouTube's on-screen auto-translate is a viewer feature, not a downloadable track; to get a translated file, download the source-language SRT and translate the text, or use the in-tool translation where offered.
Q: Can I use this to download subtitles from other sites besides YouTube? Yes — the broader free subtitle tool at JustSubtitles handles Vimeo, TikTok, TED talks, Bilibili, Rutube and friends with the same paste-URL flow, or visit the site and pick the platform from its list.
Downloading subtitles from YouTube should have been simple all along, and in hindsight it always was: the captions already exist, already are timed, already are written in a file format that every player and editor can read. What was missing was the extractor — and a browser-based downloader with a URL box is exactly that. Paste the link, pick your language, take the SRT, done in under half a minute on any device you own.
What you do afterward — proofread it, translate it, convert it for your TV, burn it into a Reel — is the same toolkit as any other subtitle file. Start at the YouTube subtitle downloader for anything living on YouTube, and the broader free subtitle tool suite covers every other platform and format on your list. Captions were always free — now they're yours, in a file you can actually use.