Audio to SRT, free in your browser
OpenSubs turns audio to SRT without sending the recording anywhere. Open an MP3, WAV or M4A on this page, Whisper transcribes the speech on your own computer, and you save a timed .srt subtitle file. It is free to use, and you do not need an account.
MP3, WAV, M4A, AAC, OGG, FLAC, Opus and WebA. No account, no watermark.
How to convert audio to SRT
The converter is the free subtitle generator on the home page. A recording goes in exactly where a video would, and the result is the same kind of subtitle file.
- On opensubs.app, press Choose a file and pick the recording, or drag it from your desktop or Finder and drop it anywhere on the page.
- Because the file has no picture, the player reads “Audio only, so there is no picture to preview.” You can still play the sound there, so you know you opened the right episode or take.
- Set the spoken language, or leave it on automatic, then press Generate from the audio. The first run downloads the speech model; after that your browser keeps it.
- Listen and read along. Retype a misheard name, change a start or end time, split a cue that runs too long or merge two short ones. Undo and redo are there if you change your mind.
- Open the Export tab and press .srt. That is MP3 to SRT, start to finish.
Audio formats you can open
The file picker on opensubs.app lists these audio extensions next to the video ones. Where each kind of file usually comes from:
| Extension | Typical source |
|---|---|
| .mp3 | Podcast episodes, voice recorders, most downloaded audio |
| .wav | Uncompressed recordings from field recorders, mixers and audio editors |
| .m4a, .aac | Phone voice memos and many recording apps |
| .ogg, .oga, .opus | Voice messages and open-source recording tools |
| .flac | Lossless archives of interviews, talks and music |
| .weba | Audio-only WebM, saved by some browser-based recorders |
What the SRT generator from audio gives you
An SRT is a numbered list of cues. Each cue holds a time range in hours, minutes, seconds and milliseconds, and the words heard in that range. Here are two cues from a recorded interview:
14 00:03:41,200 --> 00:03:44,600 We started recording in my kitchen, of all places. 15 00:03:44,900 --> 00:03:48,300 The fridge hum is on every episode from that year.
Prefer another format? The same tab saves .vtt for web players, .ass for styled subtitles, and .txt, which keeps only the words as running paragraphs with no numbers or times. If plain text is all you are after, audio to text covers that route.
When an SRT from audio is useful
- Podcast to video. If an episode goes to YouTube as a still image or a waveform clip, upload the .srt as its caption track so viewers can switch subtitles on.
- Sound recorded separately. Shot an interview with a lavalier or a field recorder? Transcribe the clean audio, then import the SRT into your editor and line it up with the picture.
- Lectures, meetings and voice notes. Timestamps on every line make it quick to find a quote and play back the exact moment it was said.
- Narration first. Record the voice-over, make subtitles from it, and keep them ready for when the edit is finished.
Subtitles in another language
Once the lines are written, open Translate and pick a target language. Every translated cue keeps the start and end time of the original, so the second SRT matches the first line for line. In desktop Chrome 138 or later the default translator runs on your device and is free. You can also use your own key for an OpenAI-compatible service, or credits on the OpenSubs backend, priced before the job runs. More on the choices: translate subtitles.
Burned-in subtitles need a video
There is no picture to draw text onto, so an audio file leaves with a subtitle file. The app says so in plain words: “Burning in needs a video file. This one is audio, so save a subtitle file below.” If you also have the video that goes with the recording, open that instead and burn the subtitles into it; that export takes a free sign-in.
How accurate the transcription is
You get Whisper’s accuracy at the model size you choose: Tiny (English only, about 40 MB), Base (about 80 MB) or Small (about 250 MB). In OpenSubs’ own test on 2026-09-09, an English clip scored a 7.7% word error rate with Base and 0.0% with Small. That clip was synthesised speech with no accent, no crosstalk and no background noise, so read the numbers as a best case; real recordings come out worse.
- Pick the language yourself when you know it, rather than leaving it on automatic.
- Try Small for a noisy room, strong accents or fast talkers. It takes roughly twice as long as Base.
- A recording that switches language is still fine: the language is checked every four seconds, and both end up in one file.
To transcribe audio to subtitles, the recording stays with you
On the default on-device settings the audio is read and transcribed inside the browser tab. The network shows the app’s own files and, the first time, the speech model; no request carries the recording. Open your browser’s Network tab while it runs if you want to see that for yourself. Cloud transcription exists, but only if you pick it: then the span you trimmed goes to the service whose key you entered.
Good to know before you start
- One file at a time. There is no batch mode, so a folder of episodes is done one recording after another.
- No file-size or length limit, but the work happens on your machine, and a two-hour recording on a laptop without a strong graphics chip takes a while. See subtitles for long videos for what to expect.
- It makes the text from what is said. It is not a tool for typing an SRT from a script.
- iOS 17 Safari cannot run the speech model; iOS 26 can.
Questions
How do I convert audio to SRT for free?
Open opensubs.app, choose or drop the MP3, WAV or M4A, press Generate from the audio, check the lines, then press .srt in the Export tab. No account and no payment are needed.
Is my audio file uploaded?
No. With the default on-device setting, Whisper transcribes the recording inside your browser tab. Only the speech model is downloaded, the first time.
Which audio formats work?
MP3, WAV, M4A, AAC, OGG, OGA, FLAC, Opus and WebA, as well as video files such as MP4, MOV, WebM and MKV.
Can I burn the subtitles into an audio file?
No. Burning in needs a video file. From an audio file you save a subtitle file instead: .srt, .vtt, .ass or .txt.
Can I get the SRT in another language?
Yes. Generate the subtitles, then translate them; each line keeps its timing. It is free on your device in desktop Chrome 138 or later, or you can use your own API key or OpenSubs credits.
Can I convert several audio files at once?
No. OpenSubs has no batch processing and works on one file at a time.