How to generate subtitles from audio
- Get the audio from your videoSpeech to Text reads audio files (MP3, WAV, M4A and other common formats). If you only have a video, export its sound first. Most video editors and many phone apps have an audio-only export option.
- Upload or paste a linkChoose the file, or switch to Paste a link if the audio is online. Anything from 5 seconds to 1 hour works.
- Pick the spoken languageChoose the language people speak in the recording. There are 43, from English and Spanish to Hindi, Arabic, Tagalog and Vietnamese.
- Set Timestamps to “Per sentence”This is the setting that makes subtitles. Each sentence gets its own start and end time, which becomes one caption.
- Transcribe, then press SRTType the 5-digit code and press Transcribe. On the finished result, press the SRT button to save the subtitle file.
What an SRT file looks like
SRT (SubRip) is the most widely accepted subtitle format. It is a plain text file with a number, a time range and the words for each caption:
100:00:00,000 --> 00:00:03,200Welcome back to the channel.200:00:03,400 --> 00:00:07,900Today I’ll show you how to repot a snake plant.
Because it is plain text, you can open it in Notepad or TextEdit to fix a name or a typo before you upload it. Keep the numbers and the arrow lines exactly as they are.
Which timestamp setting should you use?
| Setting | What you get | Use it for |
|---|---|---|
| None (plain text only) | One block of text | Show notes, blog posts, quotes, meeting notes |
| Per sentence | One timed line per sentence, plus the SRT button | YouTube subtitles, CapCut and Premiere captions, VLC |
For normal subtitles, Per sentence is the right choice. It gives captions that are easy to read at viewing speed.
Two ways to get subtitles on Speechma
Same-language subtitles
Speech to Text with sentence timestamps, then the SRT button. Best for accessibility and for viewers who watch on mute.
Subtitles in another language
AI Dubbing translates your video into a new language. Export the audio of the dubbed video, transcribe it here with sentence timestamps, and you have an SRT in that language.
Long recordings
Up to 1 hour of audio per file, enough for most podcasts, lectures and webinars.
Private
Nothing is kept on our servers. Your transcript and subtitles stay in your own browser.
Tips for cleaner subtitles
- Start with clear audio. Background music under the voice lowers accuracy. If your clip has loud music, run it through the Vocal Isolator first and transcribe the vocal track.
- Choose the right language. A Spanish video transcribed as English will come out wrong. Pick the language actually spoken.
- Proofread names and brand terms. Product names, places and slang are where AI most often slips. Fix them in the SRT with any text editor.
- Keep the audio in sync with your final cut. Export the audio from the finished edit, not from the raw footage, so the times match your published video.
- Split very long videos. Over an hour? Export the audio in parts and make one SRT per part.
Next step: put the subtitles on your video
Our guide How to add subtitles to a video shows how to upload the SRT to YouTube, import it into CapCut or Premiere Pro, play it in VLC, and when to burn captions into the picture instead.
Frequently asked questions
Is this subtitle generator really free?
Can I generate subtitles directly from a video file?
Why don’t I see the SRT button?
What languages are supported?
How long can the audio be?
Can I get subtitles in a different language than the video?
Will YouTube accept the SRT file?
Make better voiceovers, faster
- New voices and features the day they launch
- Script templates and tricks that make AI voices sound human
- Creator stories, community picks and early access
