MacWhisper — Mac download

MacWhisper turns recordings and meetings into text without sending audio anywhere. Version 15.2 adds the tools that matter for meetings.

• Real-time meeting transcripts — new in this release; system audio can be transcribed too.
• Speaker recognition — a new diarization model separates speakers.
• Folders and subfolders — keep project transcripts apart from one-off files.
• Model choice — Tiny, Small, Base, Medium, Large-V2 and Large-V3 run locally.
• Languages — 100, with automatic detection or a manual pick.
• Subtitle output — SRT and VTT, plus copying all text or only a selected section.

⌘O and ⌘R skip the import busywork

A handful of shortcuts sit just below the surface. They are not advertised loudly, but they cut friction once you know them.

• Skip the import dialog — press ⌘O to open a file directly, or ⌘R to start recording without touching the mouse.
• Jump to a segment — ⌘J opens a small search field. Type a word and the transcript scrolls there instantly. Much faster than skimming a long scroll.
• Cycle through models — hold ⌥ while clicking the model selector and the list jumps between Tiny and Large-V3 with one keystroke each. Useful when you are testing which model picks up a difficult accent best.
• Reading mode — toggled with ⌘E. It strips the segment markers and timestamps, leaving clean, flowing text. Good for pasting into a document or an email without cleanup.
• Batch export — select multiple transcripts, right-click, and choose Export. Word, PDF, HTML, SRT, and VTT all generate in one pass. No need to export each format singly.

The segment editor also rewards a closer look. Click any line, press Space to play it back, then type to correct the text. Press Tab to move to the next segment and keep editing without reaching for the mouse. This turns a rough transcript into a clean one in half the time.

Speaker separation for rough-cut show notes

The app suits people who handle spoken content regularly and care about speed more than polish.

• Podcast editors — speaker separation with the new diarization model makes it practical to produce a rough-cut transcript for show notes. Export to SRT, tidy up the speaker labels, and the subtitle file is ready.
• University students — a recorded lecture dropped into the app returns searchable notes in minutes. The reading mode and ⌘J shortcut let you find a specific concept mid-revision without replaying the audio.
• Remote team leads — real-time meeting transcripts capture decisions as they happen. Folders keep each project's calls separate. No need for a cloud service, which matters when meeting content is sensitive.
• Journalists — drag in an interview recording, let a large model handle it, then edit the segments while playing back with the Space bar. The whole workflow stays on one screen.

People who need a polished, formatted transcript with timestamps and speaker names will still do some manual cleanup. MacWhisper gives them a fast first draft, not a finished document. That trade-off feels right for the speed it brings.

Pre-installation friction points

All transcription runs locally after one download

No. Transcription runs on-device. The models download once, then everything happens locally. The real-time meeting transcript feature also processes on the Mac. Nothing is sent to a server.

Which model should I pick?

Start with Small or Base for a first pass — they are fast and surprisingly accurate on clear speech. Move to Medium or Large-V3 when you have overlapping voices, heavy accents, or background noise. The larger models need more memory and time but catch words the smaller ones miss. Model downloads range from under 100 MB to several gigabytes, so plan accordingly on a metered connection.

Does speaker separation work on older recordings?

Yes. The new diarization model in version 15.2 applies to imported files as well as live recordings. Results are better when speakers do not talk over each other. A recording with heavy crosstalk will still show mixed labels in places.

No licence check to crack or bypass

The app does not require activation. People who search for MacWhisper crack or MacWhisper torrent will find the software runs immediately after dragging it to the Applications folder. Searches like MacWhisper Pro crack or MacWhisper Pro torrent point to the same reality — no activation step exists to bypass.

Medium model stays smooth on Apple Silicon

On Apple Silicon, even Medium runs quickly and leaves the machine responsive. Intel Macs handle Tiny and Small without issue; Medium and larger models can spike CPU and fan noise for several minutes. Close heavy apps before running a long file through a large model on an Intel machine.

Drag an audio file onto the window to start, or record directly in the app.

Whisper Transcription 15.9 arrives as a maintenance update focused on stability under macOS 14 and newer. No feature changes this round.

• macOS 14 or later
• Apple Silicon or Intel 64-bit