On the Record Support

On the Record turns folders of audio and video into text transcripts, entirely on your Mac. Drop files or folders in, press Start, and get transcripts in plain text, Markdown, Word, PDF, JSON, SRT, WebVTT, CSV, or Excel — with optional speaker identification (Dialogue mode) and court-reporting layouts.

Contact

Email support@ingleworks.co with any question about On the Record.

If something went wrong, three things make it much easier to track down:

1. Your macOS version (Apple menu ▸ About This Mac).
2. The app version (On the Record ▸ About On the Record).
3. What you were doing when it happened — and, if you can, the log: ~/Library/Logs/On the Record/transcribe.log for the direct-download version. The Mac App Store version runs in a sandbox, so its log is in the On the Record folder in ~/Library/Containers/, under Data/Library/Logs/On the Record/. In Finder, choose Go ▸ Go to Folder… and type the folder's path to open it.

The log records app events and errors, never the content of your recordings or transcripts. Please don't send me recordings or transcripts unless I ask for one to reproduce a bug.

I answer every support email myself.

Requirements

• macOS 14 (Sonoma) or later
• A Mac with Apple silicon (M1 or later)
• About 2 GB of free disk space for the app; more if you download larger speech models

Frequently asked questions

Does anything leave my Mac?

No. Transcription and speaker identification run locally. The app uses the network only to download speech models, the speaker-identification model, and fonts for non-Latin PDF export — once each, and it asks first. See the On the Record privacy policy.

Why does the first transcription with a new model take longer?

Speech models other than the built-in Small model are downloaded once (78 MB to 3.1 GB depending on the model), checked against their published checksums, and then kept on your Mac. Progress is shown in the status line. Stop cancels the download between files; the file being transferred at that moment finishes in the background and is kept, so the next attempt has less to fetch. Offline, the prompt offers the built-in Small model for that run instead.

Which model should I use?

Small (built in) is the everyday default; Tiny and Base are faster and less accurate; Medium and Large v3 are more accurate and slower; Large v3 Turbo is nearly as accurate as Large v3 at several times the speed (it cannot translate). The use case you pick in Settings ▸ General (or the welcome sheet) chooses for you; you can always change it.

Where do my transcripts go?

Drop a folder and its transcripts are written to a _<Folder> Transcripts folder inside it. Files added individually are saved next to the media (or you are asked where). Turn on "Ask where to save transcripts" in Settings ▸ Output to be asked every time.

What is Dialogue mode?

It adds speaker identification, so the transcript reads as a conversation. Set the number of speakers if you know it (it noticeably improves accuracy), then name the speakers after transcription — you can play a sample of each voice.

How do I free up disk space?

Settings ▸ Advanced ▸ Remove Downloaded Models deletes downloaded speech models; they download again if you pick them later. The built-in Small model is never removed.

A PDF warned that a font was not downloaded.

PDFs containing Chinese, Japanese, Korean, Arabic, or Hebrew need a Unicode font, which the app downloads once with your permission. If you declined, allow it in the prompt next time and re-export.

Which media formats are supported?

Video: MP4, MOV, AVI, MKV, M4V, WebM. Audio: MP3, WAV, M4A, FLAC, AAC, OGG, WMA, AIFF.

A file failed to transcribe.

Hover the failed row for the reason. Common causes: the file has no audio track, is DRM-protected, or is not a media file despite its extension. Retry from the row's context menu after fixing the file.

On the Record is made by Ingleworks LLC, Phoenix, Arizona.