What this YouTube transcript tool actually does
YouTube already stores a text track for a huge share of its videos. Some are written by the creator, some are uploaded by a captioning service, and many are generated automatically by YouTube's speech recognition. The problem is that YouTube only lets you read that text inside a narrow side panel, one line at a time, with no clean way to copy it, search it properly or save it as a file.
PullSubtitle reads the public caption track for the video you paste and hands it back to you as a proper document. You get every line with its start time, the full plain-text version, the language of the track, and the video's details such as title, channel, publish date, length, views, likes and comment count. Nothing is re-recorded and nothing is guessed: the words are the ones attached to the video on YouTube.
Because we work from the published captions, length is not a problem. A ninety-second Short and a four-hour podcast episode are handled the same way, and a long lecture comes back as quickly as a short clip. The only requirement is that the video is public (or unlisted with the link) and that YouTube has a caption track for it.
How to get the transcript of a YouTube video, step by step
If you have never pulled a transcript before, the process takes about ten seconds:
- Open the video on YouTube in your browser or the app.
- Tap Share and choose Copy link, or copy the URL from the address bar.
- Paste it into the box at the top of this page.
- If the video has captions in several languages, pick the one you want from the Language menu. Otherwise leave it on Auto.
- Press Get transcript. The text appears below with the video's details on the left.
- Toggle Timestamps on or off, search for a keyword, and copy or download the result.
Links in any of these shapes are accepted: youtube.com/watch?v=..., youtu.be/..., youtube.com/shorts/..., m.youtube.com links from your phone, and links that carry extra parameters such as a start time or playlist ID. We strip what we don't need.
Getting the transcript on a phone
The same steps work on iPhone and Android. In the YouTube app, tap Share, then Copy link, open PullSubtitle in your mobile browser and use the Paste button inside the input. The transcript view is built to scroll comfortably on a small screen, and Copy puts the whole text on your clipboard in one tap.
Why not just use YouTube's built-in “Show transcript”?
YouTube's own transcript panel is useful for a quick look, and you should use it if that is all you need. It falls short the moment you want to do something with the text:
- Copying is clumsy. You have to drag-select inside a scrolling panel, and you usually end up with timestamps glued to every line.
- No files. There is no button to save SRT subtitles, a spreadsheet or a document.
- Search is limited. You can't easily count how often a phrase is said or jump between matches.
- It isn't always visible. The panel is hidden on some layouts and missing in parts of the mobile app.
- One video at a time. There is no way to work through a list of videos.
PullSubtitle gives you the same underlying text with a clean copy button, timestamp control, in-transcript search with match counts, six export formats and, on paid plans, batch processing for up to fifty links in one run.
Export formats and when to use each one
Every format is generated from the same caption data, so timings line up across all of them.
- TXT — plain text, optionally with [m:ss] markers. Best for reading, quoting, pasting into a doc or feeding into an AI assistant.
- SRT — the most widely supported subtitle file. Import it into Premiere Pro, DaVinci Resolve, Final Cut, CapCut or upload it to another platform.
- VTT — the web subtitle standard used by HTML5 players, Vimeo and most video hosts.
- CSV — one row per caption with start and end times in both clock format and seconds. Open it in Excel or Google Sheets to analyse pacing or tag segments.
- JSON — structured data including the video metadata and every segment, ready for scripts, databases and research pipelines.
- Markdown — a tidy document with the title, channel, date, duration and a timestamped transcript. Drop it straight into Notion, Obsidian or a GitHub wiki.
The free plan includes TXT and SRT. Starter and Pro unlock all six.
What people use YouTube transcripts for
Turning videos into written content
Creators and marketers take a transcript, tighten it up and publish it as a blog post, newsletter, LinkedIn article or Twitter thread. A twenty-minute video is roughly three thousand words of raw material, which is far faster to edit than to write from scratch.
Studying and note-taking
Students paste lecture links to get searchable notes, find the exact moment a concept was explained, and build revision sheets. Language learners read along with native speakers and look up words they missed.
Research and journalism
Researchers quote speakers accurately, check what was actually said in interviews and hearings, and code long recordings for qualitative analysis. The CSV and JSON exports keep timings attached so every quote can be traced back to the second it was said.
Subtitles and accessibility
Editors download SRT or VTT to subtitle a re-edit, a translated version or a clip posted on another platform. Viewers who are deaf or hard of hearing, or who simply prefer reading, get a version they can follow at their own pace.
SEO and content audits
Agencies pull transcripts to see which keywords a competitor's video actually covers, to write chapters and descriptions, and to make sure their own videos say what the title promises.
Summaries with AI
A transcript is the input most AI tools need. Copy it into your assistant of choice and ask for a summary, key takeaways, quiz questions or a translation. Using the real captions instead of a guess from the title gives far better results.
Choosing a caption language
Many YouTube videos carry more than one caption track: the original language, an auto-generated track, and translations uploaded by the creator. When you leave the Language menu on Auto / original, you get the default track YouTube would show. When you choose a language, we ask for that track specifically.
A few practical notes:
- If the language you pick doesn't exist for that video, you will get the closest available track or a no-captions message. Empty results don't use credits.
- The video card on the left lists the caption tracks we found, so you can see which languages are available before trying again.
- Auto-generated captions are usually accurate for clear speech in widely spoken languages, and weaker with heavy accents, music, crosstalk or specialist vocabulary. Creator-uploaded tracks are normally the cleanest.
Video details that come with every transcript
Alongside the text, the left-hand panel shows what the video is: title, channel name and handle, upload date, duration and thumbnail. On paid plans you also get likes, comment count, an engagement rate calculated from views, the video's keywords and tags, its YouTube category, the number of chapters and the full list of caption tracks.
That context matters when you are collecting transcripts for research or competitive analysis. When you export JSON or Markdown, the metadata travels with the text so you never lose track of where a quote came from.
Pro subscribers can also open the Comments tab to load up to a hundred top comments for the video and download them as a CSV, which is handy for audience research and for spotting the questions viewers keep asking.
When a YouTube video has no transcript
Occasionally PullSubtitle will tell you no transcript was found. The usual reasons are:
- Captions are turned off. Creators can disable automatic captions, and some never upload their own.
- The video is too new. YouTube generates automatic captions after upload; very fresh videos sometimes don't have them yet. Try again a little later.
- Music or silence. Videos with no speech, or with speech buried under loud music, often have no usable track.
- Unsupported language. Automatic captions exist for many but not all spoken languages.
- Private, age-restricted or removed videos. We can only read what is publicly available.
Empty results never use credits. Only successful transcripts do.
Accuracy, privacy and fair use
The words you receive are YouTube's own caption track, so accuracy depends on how that track was made. Uploaded captions are typically near-perfect; automatic captions can miss names, jargon and punctuation. We clean up markup, merge repeated lines and normalise spacing, but we don't rewrite what was said.
We don't need your YouTube account, we never ask you to log in to Google to read captions, and we don't download or store the video itself. If you are signed in, your transcripts are kept in your private history for as long as your plan allows (3 days on Free, 30 days on Starter, 90 days on Pro) and you can delete any of them at any time.
Transcripts are for personal study, research, accessibility and commentary. If you plan to republish someone else's words at length, get the creator's permission or make sure your use falls within fair use where you live.