How to transcribe a YouTube video into text and captions

The steps work for videos you made and videos whose owner has given you the rights. Transcription is multilingual, and a sentence that switches language keeps each word in the script it was spoken in.

You can transcribe your own uploads and any video you have the rights to. A video someone else posted belongs to its owner, and so do its words. Ask the owner before you transcribe it, and again before you publish anything made from the transcript.

YouTube Studio lets you download your own uploads. YouTube's help page says a video with a copyright or Community Guidelines strike, or one that uses a preapproved audio track, cannot be downloaded.

Start from the best copy you have, which is usually the file your video editor exported.

1. 01Export the audio from your editorIn the project you edited, export the soundtrack alone as an M4A or MP3 file. Export it from the final cut, after every trim, and the timing will match the video you published.
2. 02Or download the video from YouTube StudioSign in to YouTube Studio, open Content, select the video, open its menu, and choose Download. YouTube's help page says the file is an MP4 at 720p or 360p, depending on the size of the video.
3. 03Take the sound out of the MP4Open the MP4 in a video editor or an audio app and export the sound as M4A or MP3. An audio file is a fraction of the size of the video, and one file can run up to 120 minutes and 95 MB.

1. 01Upload the audioOpen the Transcribe tool and drop the file in. The cost appears before you start.
2. 02Choose Written numbersWritten gives digits and units, which is what captions want. Spoken keeps every number in words, exactly as it was said.
3. 03Choose the styleClean drops fillers and false starts and reads more easily on screen. Verbatim keeps every word, for a video where the way something was said matters.
4. 04Add your vocabularyList your channel's name, your products, your guests, and any jargon, up to 50 terms. They come back spelled the way you wrote them.
5. 05Turn on Speakers for a conversationFor an interview or a show with two hosts, Speakers returns each turn with a number, and the caption file starts a new line when the speaker changes.
6. 06Pick SRT or VTT and transcribeUnder Subtitles, choose SRT or VTT, run the transcription, and download the subtitle file. Copy the transcript as well, for the description and anything else you write from it.

1. 01Open SubtitlesSign in to YouTube Studio, choose Subtitles from the left menu, and select the video.
2. 02Add the languageClick Add language and choose the language the video is spoken in. Then click Add under Subtitles.
3. 03Upload the fileChoose Upload file, then With timing, because an SRT or VTT file carries its own timing. Choose the file and select Save.

YouTube accepts SubRip (.srt) and WebVTT (.vtt) files, among the formats its help page lists. Play the first minute with captions on and check that each line appears as it is spoken. YouTube's own steps cover the other ways to add captions.

YouTube can caption a video automatically with speech recognition. Its help page on automatic captions says their quality may vary, and it asks creators to review them and fix whatever was transcribed wrong.

Your own file puts that review before the video goes live. The vocabulary spells the names, you read the transcript once, and the same text then feeds the description, a blog post, and a translation.

• Names of people, brands, and places, which viewers notice first.
• Every figure, from prices to dates.
• Lines that run too wide for a phone screen. Line length in the Transcribe tool sets how wide each caption runs.
• The first and last captions, which show whether the timing still matches after your edits.

Word timestamps make the check quick. Turn them on, find a doubtful word, play the second it was said, and correct it in the file.

• Write the video's description from it, opening with what the video covers.
• Edit it into a blog post or a newsletter, which search engines can read.
• Pull quotes for Shorts and posts, with the second each one was said.
• Use it as the script for a dubbed version voiced with text to speech. The dubbing page shows the workflow.

Transcription is priced by the length of the audio. Paxa STT Lite costs 8.33 credits per minute. An hour is 500 credits, or $0.50.

A new account's 100 free credits cover about 12 minutes of audio, with no card. They are for personal use and evaluation. Buy any plan or credit pack, and the account's transcripts and captions are yours to use commercially, including those made before the purchase. A monetized channel belongs on a paid account, and the Terms state the rule.

• Upload a video's audio to the Transcribe tool with the free credits.
• See a subtitle file cut from a real transcript on the subtitle generator page.
• Read how to transcribe a podcast for a show with two or more hosts.
• Compare transcription services on the speech to text API comparison.

Can I transcribe someone else's YouTube video?#

Only with the owner's permission. A video and the words in it belong to whoever made it. Ask first, and credit the owner in anything you publish from the transcript.

Can I upload the video file itself?#

Export the audio first. The Transcribe tool reads audio files such as M4A and MP3, which are far smaller than the video. The captions still line up with the video, because the timing is the same.

How long can the video be?#

One file can run up to 120 minutes and 95 MB. Split a longer video's audio at a natural break and transcribe each part. Each part's captions start from zero, and a subtitle editor shifts the later parts by the length of the parts before them.

Should I download SRT or VTT?#

YouTube accepts both. SRT is the older and plainer format, and VTT is the caption format built for the web. Choose whichever your editor and your other platforms read.