Transcription
Vida transcribes recordings on your Mac with on-device speech recognition. Recording and transcript content never leaves your Mac.
Free transcriptions and Vida Pro
- Vida Free includes 5 complete, full-quality transcriptions of recordings made in Vida, including speaker identification, language selection, and every export format. They do not renew.
- Before using one, Vida asks Use a Free Transcription? — This uses 1 of your 5 free transcriptions. Choose Use It or Cancel.
- Only a transcription that completes counts. Failed and cancelled ones cost nothing, and transcribing the same recording again is always free.
- When all 5 are used, choosing Transcribe opens the Vida Pro sheet. Transcripts you already have stay yours.
- Vida Pro removes the limit and adds Auto-Transcribe, glossary corrections, and importing files to transcribe.
Transcribe a recording
- On the completion screen, click Transcribe (it shows your current Spoken language). Or, in Recent, click a row's transcribe icon.
- Follow the progress on the completion screen, in Recent, or in the Transcript tab.
- When it finishes, click Open Transcript.
The first transcription downloads the speech model, so it takes longer. Keep Vida open and the Mac awake.
Choose the spoken language
Set Settings → Transcription → Transcription Language to the language spoken in your recordings: Auto-detect (default), Spanish, English, French, German, Portuguese, Italian, Japanese, Russian, Chinese, or Korean. This is separate from Vida's interface language.
With Auto-detect, Vida listens to part of the recording first to guess the language. That takes extra time and can guess wrong on short, quiet, or multilingual audio. While Auto-detect is selected, Vida shows a tip with Choose Language and Not Now.
Choose the speech model
Settings → Transcription → Speech Model offers Tiny, Base, Small, and Large v3 Turbo. Larger models are more accurate but take longer and use more memory. Under the picker, Vida describes the selected model and its download size (about 74 MB, 140 MB, 483 MB, and 3 GB respectively).
On a new installation, Vida picks the model your Mac's memory can comfortably run — Large v3 Turbo with 16 GB or more, Small with 8–16 GB, and Base below that — and marks it (recommended).
Identify speakers
Settings → Transcription → Identify Speakers (on by default) labels each transcript segment by speaker. The speaker model downloads once (about 12 MB); its status shows under Speaker Model, with a Download button if it isn't installed yet. If speakers can't be identified, the transcript is still saved, without labels, and Vida says so.
Merge speaker segments combines consecutive entries from the same speaker into readable paragraphs with one time range. It needs Identify Speakers on.
Speaker labels are automatic and can be wrong when people talk over each other.
The Transcript tab
The Transcript tab shows transcriptions that are queued, running, paused, or failed, plus those finished in the last 24 hours. Older finished transcripts are in Recent. Processing continues in the background while you use the rest of Vida.
Each job shows its status — for example Loading speech model, Detecting spoken language, Transcribing part 2 of 5, Identifying speakers — with a progress bar and an estimated time remaining. A finished job reads Completed in followed by how long the work itself took.
Actions on each job:
- Pause stops after the current part and keeps everything transcribed so far. Resume continues from where it stopped, even after Vida has been reopened. You can also pause from the completion screen and resume from Recent.
- Cancel stops the job.
- Retry runs a failed job again. If the recording can't be found, Locate… lets you point Vida at it.
- Open Transcript opens a finished transcript.
- The trash icon deletes the transcript (see below).
When a recording starts, Vida pauses transcription after the current part (Paused to prioritize recording) and resumes automatically when the recording ends.
Read, copy, and export a transcript
Open Transcript shows each segment with its timestamp and, if identified, its speaker. From the toolbar:
- Copy copies the transcript as text.
- Export saves it as Markdown, Text, SRT subtitles, or WebVTT captions.
- Re-transcribe creates a new transcript with your current settings and replaces this one when it finishes. The recording itself is not changed.
- Open Recording plays the source recording.
- Delete Transcription permanently deletes the transcript. The recording remains and can be transcribed again. This cannot be undone.
- Close closes the transcript.
Transcripts stay readable, copyable, and exportable in every format even if your Vida Pro access later ends.
Auto-Transcribe (Pro)
Settings → Transcription → Auto-Transcribe queues each recording for transcription as soon as it finishes, and each file you import. It requires Vida Pro because it cannot ask before using a free transcription.
Glossary
The glossary fixes words the engine consistently mishears, such as names, products, or acronyms.
- Open Settings → Transcription → Glossary and click Manage….
- Type the word as it is misheard (Misheard as…) and its correct spelling (Correct spelling…), then add it.
- Click Close.
The glossary is part of Vida Pro: adding corrections and applying them to new transcriptions require it. Without Vida Pro you can still open the glossary to review, edit, or remove the entries you already have.
Where transcripts are saved
Each transcript is saved next to its recording, in two subfolders of the recording's folder: json/ (the transcript data) and transcriptions/ (a Markdown copy). If Vida can't write in the recording's folder, it uses your save folder instead.