RoostTX - Transcribe Audio

RoostTXTranscribe Audio

macOS

Drop audio. Get text. On-device transcription.
No internet. No uploads.
Nothing leaves your Mac.

⌘ wifi 9:41 AM

Why I made this.

For a few years now I have been using a voice recorder to document story and play ideas, as well as any other big thoughts I might be working through. Needing to transcribe them took me all over the internet, including AI-created pages where you uploaded audio- something I found frustrating and a tad worrisome.

I wanted to streamline the process without it breaking my bank account.

This app does two things: transcribe audio, and make subtitles. That is it. That is all. Knowing RoostTX in the menu bar taking up no space- waiting for my next bit of inspiration to be dropped into the black hole- gives me so much joy. Audio to text in no time. Now I am the one holding things up.

Drag in. Walk away.

Click the menu-bar icon, drop your files into the panel, and RoostTX handles the rest. Individual files, whole folders, or a mix — it walks folders recursively and processes everything it finds.

Files being dragged onto the RoostTX panel

Individual files	Drop any number of audio or video files at once
Folders	Drop a folder and the app finds every audio and video file inside, including subfolders
Spring-loaded icon	Drag files directly to the menu-bar icon — hold for a moment and the panel springs open automatically
Pin	Keep the panel open while dragging from Finder with the pushpin in the title bar

Two files per transcription.

A finished transcription in the RoostTX panel

Every transcribed file produces two outputs, written to your chosen folder automatically.

Plain text	The spoken words, nothing else. ready to copy, edit, or search
Text with timecode	Pick your preferred timing and RoostTX will track

Subtitles? We can do that.

Select the SRT button in the title bar and done. RoostTX tracks the timecode when dialogue is spoken. Names, and formatted so you can immediately drop it into any video playback program.

Private by design.

Transcription runs entirely on-device using Apple's built-in speech engine. Your audio is never sent to Apple or anywhere else. No microphone permission required — the app reads files from disk only.

The first time you transcribe in a language, macOS may download a language model for that locale. That one-time download requires an internet connection; all subsequent transcriptions work offline.

A few things worth knowing.

Requires macOS 26 (Tahoe) or later — the on-device transcription engine was introduced in macOS 26.
Supports any audio or video format macOS can read: MP3, M4A, WAV, AIFF, MP4, MOV, and more.
Files with the same name are never overwritten — duplicates are renamed automatically (recording_02.txt, etc.).
Empty or unreadable files are skipped; the batch continues and shows a skipped count when done.

The empty RoostTX panel waiting for files — Drop files here

A batch of files being transcribed — Batch