Turn audio, podcasts and recordings into subtitles on your Mac

You have a recording, a podcast episode, an interview or a voice memo, and you want timed subtitles, or text you can copy and search, without sending the audio to someone else's server. Twinsubs is not only for video: drop in an mp3, m4a or wav and you get an .srt beside it just the same. If the episode is in a language you do not understand, the same run can add a translation so you can read along as you listen.

What it can process

• Audio: mp3, m4a, wav, aac, aiff, flac and caf.
• The sound of a video: mp4, mov and other video files work too, and it is their sound that gets recognized.
• Many at once: drop a batch in and it works through the queue in turn.

Three steps

1. Drop it inDrag the audio file into the window.
2. Wait a momentThe source language is detected for you, or you can set it by hand; you choose the target language. To get the original text only, pick the same language as the target. The first launch needs a connection to download a speech model once (under 1 GB).
3. Find the subtitles beside the audioThe same folder gains an .srt: a single-language file such as Podcast.en.srt, and for foreign-language audio a bilingual one such as Podcast.en-zh.srt (target first, source second).

What you can do with the SRT

• Use it as subtitles: for a video podcast or an audio show with pictures, import the SRT into an editor, as described in import an SRT.
• Turn it into a plain-text transcript: an SRT is plain text, so removing the numbers and timecodes leaves a transcript. The method is in the section on converting to plain text in check and edit SRT subtitles.
• Read along with a foreign-language episode: the translation sits above the original, so a sentence you missed is clear at a glance and the original stays there to compare.
• Find a spot by keyword: search the SRT for a phrase, and the timecode beside it is where it sits in the audio.

In file mode Twinsubs writes SRT. It does not export txt directly or a transcript with speaker names. For plain text, convert the SRT as described above.

Which audio suits it

• Podcasts and interviews: your own show, or a foreign-language episode.
• Course and lecture recordings: turn them into subtitles or notes; see subtitles for online courses and lectures.
• Voice memos and interview material: content you would rather not upload stays on your Mac.
• Listening material in a foreign language: read the two lines side by side.

When a recording contains other people's voices, make sure you are allowed to process it, and tell them first where that is appropriate.

Limits, stated up front

• No speaker separation: a conversation comes out as continuous subtitles, with no marking of who said what.
• Accuracy: loud background music, people talking over each other, heavy accents and technical terms make mistakes more likely. Check numbers and names against the audio.
• Not live: this processes audio files that already exist. To translate while something plays, see live translate videos and streams.
• No lyric alignment: it recognizes speech.
• Allowance: the free version processes the first 10 minutes of each file; the purchase removes the limit.
• System: a Mac with macOS 26 or later, Apple silicon.

Privacy: the audio stays on your Mac

Recognition and translation both run on your Mac. There is no backend, so the audio never passes through a server. Only the first launch needs a connection, to download the speech model once; translation language packs come from macOS and work offline afterwards. See the privacy policy.

FAQ

Updated 2026-10-12. Back to the Twinsubs home page