How to make SRT subtitles for a YouTube video

You can make an SRT subtitle file from a video's speech, check it for mistakes, then upload it to YouTube. Automatic transcription gives you a useful first draft, but you should review the words and timing before viewers rely on it.

The workflow is simple: transcribe the video or its audio, save the result as an SRT file, and upload that file in YouTube Studio. You can also use a separate audio file if that is easier to work with.

Create the SRT file and upload it

1. Open the Speech to Text tool. Use Speech to Text and select the video or audio file from your device. The file is processed in your browser and is not uploaded.
2. Choose the spoken language. Select the language used in the recording. If the video switches languages, check the transcript carefully because a single language choice may not recognize every passage correctly.
3. Transcribe the recording. Start the file transcription and wait for the text and subtitle timings. A clear recording with one speaker is usually easier to transcribe than overlapping voices or speech under loud music.
4. Review the transcript against the video. Play the recording and follow along with the generated text. Correct names, technical terms, punctuation, and any words the transcription missed. Listen especially closely where speakers talk quickly or the audio is noisy.
5. Save the result as SRT. Choose SRT as the output format and save the file. Keep the .srt extension so YouTube recognizes it as a subtitle file.
6. Upload it to YouTube Studio. Open the video's subtitle settings, add the relevant language, and choose the option to upload a file. Select the SRT file and choose the option that says it already has timing. Review the captions in the preview before publishing.

YouTube Studio labels and layouts can change, but the important choice is to upload a subtitle file with timing. Do not choose an option that asks YouTube to create timings for text that already contains them.

Check that the SRT timing and format look right

SRT is a plain-text format. Each caption entry, called a cue, has a number, a start and end time, the caption text, and a blank line before the next cue. A timestamp looks like 00:00:01,200 --> 00:00:03,500. The comma separates seconds from milliseconds.

Open the saved file in a plain-text editor if you want to inspect it. The first cue usually starts at the beginning of the video, and the later cues should appear in the same order as the speech. Numbers should increase in sequence. The caption should appear while the words are spoken, not several seconds before or after.

Keep each cue short enough to read while watching. A sentence can be split at a natural pause, but avoid breaking apart a name or phrase in a way that changes its meaning. If a line is too long, divide it across cues or lines without changing the spoken words.

Review the words before publishing

Correct names and specialized terms

Automatic transcription can mishear names, place names, acronyms, and uncommon terms. Search the SRT for words you know may be difficult, then replay those moments. Check capitalization and punctuation too, especially if a mistake could change the meaning.

Check speech that is hard to hear

Background music, room echo, noise, and speakers talking over one another can produce missing or incorrect words. Listen to the original recording rather than guessing from the sentence around the mistake. If a word cannot be understood, do not replace it with a confident guess.

Make sure the captions are readable

Captions should follow the speaker's phrasing and leave viewers enough time to read. Avoid one cue for a whole long paragraph, and avoid making a new cue for every word. Check that the text does not cover important on-screen details when YouTube displays it.

Fix common SRT upload problems

YouTube does not accept the file

Check that the file ends in .srt and is saved as plain text, not as a document from a word processor. If you open and resave it in a text editor, choose UTF-8 encoding when that option is available. Confirm that each cue has a number, a timestamp line, caption text, and a blank line before the next cue.

The captions are out of sync

Confirm that the SRT was made from the same version of the video you uploaded. If the video has an opening title, removed section, or different audio edit, its captions may no longer line up. Check the beginning, middle, and end in YouTube's preview. If the shift changes over time, review the cue timings throughout the file instead of adjusting only the first caption.

The transcript has many errors

Check that you selected the right spoken language and used the clearest available recording. If the video has separate audio and video files, make sure they are from the same edit. Correct important errors manually before publishing, and do not treat automatic transcription as a substitute for listening.

The captions show strange characters

This can happen if the file was saved with an incompatible text encoding. Reopen the SRT in a plain-text editor, save it as UTF-8 if possible, and upload the newly saved file. Also check that the editor did not change the timestamp punctuation or remove blank lines between cues.

Keep the original video and the edited SRT together until you have checked the uploaded captions from start to finish. The transcription file stays on your device while you create it, but uploading the SRT to YouTube shares those captions with the video according to its visibility settings.