How to Transcribe a YouTube Video for Free

The quickest free way to transcribe a YouTube video is to copy the transcript YouTube already made: open the video in a browser, click Show transcript in the description, and copy the text (YouTube Help). That only works when the video has captions. If it has none, or you need a clean file with timings, you have three more free options: download the captions of your own video from YouTube Studio, run OpenAI's open-source Whisper model on a file you have, or use the free plan of a transcription tool. Each is covered below, with what it costs you in time and what the result looks like.

Steps checked against YouTube's help pages and the Whisper repository on September 30, 2026.

The four free methods at a glance

Method	Works for	What you get	Effort
Copy YouTube's transcript	Any video with captions	Plain text with timestamps, pasted wherever you like	A minute
Download captions in YouTube Studio	Your own videos	A caption file with timings	A few minutes
Whisper on your computer	Audio or video files you have	TXT, SRT, VTT and other formats	An hour to set up the first time
A transcription tool's free plan	Links or uploads, within the plan's limits	An editable transcript and caption files	A few minutes

Method 1: Copy the transcript YouTube already made

YouTube shows a transcript for any video that has captions, including its automatic captions.

Open the video on youtube.com.
Expand the description and click Show transcript.
The transcript opens and scrolls along as the video plays. Click any line to jump to that moment.
Select the text in the panel and copy it into a document.

On the YouTube app the steps are the same: open the description and tap Show transcript.

What you get is plain text with a timestamp on each line. To remove the timestamps, paste into a spreadsheet or text editor and delete the timestamp lines, or use one of the methods below that exports clean text.

Limits: no captions, no transcript. And automatic captions can be wrong. YouTube itself warns that they "might misrepresent the spoken content due to mispronunciations, accents, dialects, or background noise" (YouTube Help). Automatic captions are available in a fixed list of languages, and YouTube may not generate them at all if the video is too long, has poor audio, starts with a long silence, or has several people talking over each other.

For more on finding and copying this panel, including on your phone, see how to get the transcript of a YouTube video.

Method 2: Download captions for your own video

If the video is on your channel, YouTube Studio lets you download the caption track as a file.

Sign in to YouTube Studio and select Subtitles in the left menu.
Click the video.
Under Subtitles, click Edit next to the language you want.
Fix any errors in the editor first. This is the best moment to correct names and jargon.
Click Options and select Download subtitles. Your browser downloads a file with the caption track.

(YouTube Help)

Tip: if you already have a written script or a transcript without timings, YouTube can time it for you. In YouTube Studio's Subtitles page, select the video, click Add language, then under Subtitles click Add and choose Auto-sync. Paste the text or upload a transcript file. YouTube syncs it to the video, but only for supported languages, and it doesn't recommend auto-sync for videos over an hour or with poor audio (YouTube Help).

Method 3: Run Whisper on your own computer

Whisper is OpenAI's open-source speech recognition model. It runs on your own machine, so there's no usage cap and no fee, and the audio never leaves your computer. The trade-off is setup.

You need Python (the repository says the code is expected to work with Python 3.8 to 3.11), the ffmpeg command-line tool, and ideally a graphics card. Then:

1pip install -U openai-whisper
2whisper episode.mp4 --model turbo --output_format srt

Leave out --output_format and Whisper writes every format it supports: TXT, VTT, SRT, TSV and JSON. Add --language English (or another language) if auto-detection guesses wrong; by default Whisper detects the language from the first 30 seconds (Whisper repository).

Choose a model to fit your hardware. Per the repository's table, turbo needs about 6 GB of video memory, small about 2 GB and tiny or base about 1 GB. Smaller models are faster and less accurate. Note that turbo isn't trained for translation, so use medium or large with --task translate if you want an English translation of non-English speech.

Whisper needs a file. That means your own recording, or a video you have the rights to download; YouTube's Terms of Service don't allow downloading other people's videos without permission (YouTube Terms of Service). For a deeper look at the model, see our guide to OpenAI Whisper.

Method 4: Use a transcription tool's free plan

Web tools take a YouTube link or an upload, transcribe it, and give you an editor to fix mistakes before you export. Many have a free plan with limits on length or usage, so check what the plan covers before you start on a long video.

Disclosure: Exemplary AI is our product. On the YouTube transcription page you paste a link (or upload the file), choose the language, and get a transcript you can edit, in any of 99 languages. You can download it as DOCX, TXT, SRT or VTT. The Free plan comes with a one-time allowance of 500 credits and takes videos up to 30 minutes long; see pricing for more. Other options are compared in our best transcription software roundup.

Which method should you use?

You just want to read or quote it: Method 1.
It's your video and you want accurate captions on YouTube too: Method 2. Fixing the captions in Studio improves your video and gives you a clean transcript in one go.
You have many files, or the audio can't go to a third-party service: Method 3.
The video has no captions, or you want an editable transcript and caption files without installing anything: Method 4.

Clean up the transcript

Whatever method you use, spend five minutes on these before you publish or share the text:

Names and terms. Search for every person, product and place name, and fix the spellings.
Numbers. Check figures, dates and prices against the audio.
Speakers. Add labels if more than one person talks.
Paragraphs. Break the text at topic changes so it's readable.

If you want a summary rather than a full transcript, paste the text into ChatGPT with one of the prompts in how to summarize a YouTube video with ChatGPT.

FAQ

Can I transcribe a YouTube video for free?

Yes. Copying YouTube's own transcript is free, downloading captions of your own videos in YouTube Studio is free, and Whisper is open source and free to run on your own computer. Many transcription tools also have free plans with usage limits.

How do I transcribe a YouTube video that has no captions?

Run a speech-to-text tool on the audio. If it's your video, download it from YouTube Studio and use Whisper or a transcription tool. If it isn't, some tools accept a YouTube link, but you should still only transcribe videos you have the right to use.

How accurate are YouTube's automatic transcripts?

YouTube doesn't publish an accuracy figure. It says automatic captions can misrepresent speech because of accents, dialects, mispronunciations or background noise, and recommends reviewing and editing them (YouTube Help).

Can I get the transcript as an SRT file?

For your own videos, download the caption track from YouTube Studio. Whisper can write SRT directly with --output_format srt, and transcription tools usually export SRT. See how to make an SRT file for the format itself.