Private by design. Powerful by default.

Every voice.
Clearly captured.

Transcribe recordings, live meetings and video on your Mac - with local AI, speaker detection, editable text and professional exports.

● Download full appSee it in action ↓
Apple SiliconmacOS 14+No API credits
TranscribeIt showing an editable multi-speaker transcript
Processing stays on your Mac
12 speakers detected

Audio files • Live meetings • Video • Multiple speakers • Local AI

A complete local workflow

From raw recording to finished document.

Explore the real interface. Every screenshot comes from the full app.

Listen, inspect, correct

Stay in sync with the source.

Click any timecode to hear the exact moment, slow playback down, edit the text inline and rename every speaker. For video, the player and transcript follow each other in both directions.

1 / 4
Editable multi-speaker transcript with synchronized audio controls
Everything you need

One app. The whole workflow.

Capture, verify, improve and deliver - without moving sensitive recordings between services.

◎

Audio & video import

Drop in common audio or movie files. Transcribe several files in sequence, with a local media copy saved for reliable processing.

◉

Live meetings

Record from the microphone, see live progress, press Stop and let the app complete a higher-quality final pass.

文

Many languages

Use a known source language for stability or let Whisper detect it automatically. Swedish, English and German are first-class workflows.

⌁

Apple Speech or Whisper

Start with macOS speech recognition or download a local Whisper model for consistent multilingual transcription.

A model library you control

Choose compact. Choose powerful. Choose local.

The app is lightweight. Download only the engines you want, use them without API credits, and remove them whenever you need the disk space back.

Available locally

Local speech-to-text. Pick speed, size or maximum accuracy.

01
TinyFastest · multilingual
~75 MB
Ready
02
BaseFast · multilingual
~145 MB
Ready
03
SmallBalanced
~500 MB
Ready
04
Large v3 TurboRecommended · high accuracy
~626 MB
Ready
05
Large v3 Turbo fullMaximum quality
~1.5 GB
Ready
06
English variantsTiny, Base and Small
75-500 MB
Ready

Model sizes are approximate. Chatterbox and its local engine require several GB. Chatterbox is strongest in English; Swedish and other languages remain experimental.

Privacy is a feature

Your recording is not a cloud prompt.

Whisper, Gemma, SpeakerKit, Kokoro and Chatterbox run on your Mac after download. No API key. No credit meter. No routine upload of your transcripts.

  • 01 Imported media is copied into local app storage for reliable review.
  • 02 AI summaries and proofreading can run through downloaded Gemma models.
  • 03 Original audio and corrected speech exports are generated locally.
Deliver it your way

From transcript to whatever comes next.

TXTRTFDOCXSRTVTTJSONPDFLaTeXM4A
Transcript

Editable documents

Text, rich text and Word files that open cleanly in the tools you already use.

Video

Timed subtitles

SRT and WebVTT with segment timing, plus JSON for structured workflows.

AI result

Rendered and portable

Word, PDF or complete LaTeX for polished reports and Overleaf projects.

Audio

Original or regenerated

Archive the original M4A or create multi-voice audio from your corrected text.

Built differently

Made with ChatGPT Work. On a Mac.

TranscribeIt Local was built through an extended, hands-on collaboration with ChatGPT Work on macOS - from the first native Swift interface to live transcription, local models, audio synthesis, exports, testing and release packaging.

ChatGPT helped build the product. It is not required to run the app, and the local workflows do not consume ChatGPT API credits.

01Describe the needNative transcription, privacy, languages
02Build and testSwift, real media, real meetings
03Iterate togetherSpeakers, AI, video, voices, export
04Ship a Mac appSigned full edition · version 1.0
Good to know

Before you download.

What exactly is included in the download?+

The full TranscribeIt Local 1.0 app with every feature enabled. The DMG is small because large Whisper, Gemma, SpeakerKit, Kokoro and Chatterbox models are downloaded on demand inside the app.

Which Mac do I need?+

An Apple Silicon Mac running macOS 14 or later. More memory helps with the largest transcription, AI and neural voice models.

Does it work without internet?+

Yes, after your selected models have been downloaded. Internet is needed for initial model downloads. Apple system features depend on what your macOS version supports locally.

Can I transcribe film and follow the picture?+

Yes. Imported video can remain visible while you review. Text seeks the movie, and the current movie position can jump back to the corresponding transcript row.

Can I correct the transcript before export?+

Yes. Text rows, speaker assignments, speaker names and transcript titles are all editable. You can also delete incorrect segments and run selective AI proofreading.

How do I install it?+

Download the DMG, open it and drag TranscribeIt Local to Applications. On first use, allow microphone and speech-recognition access when macOS asks.

TranscribeIt Local app icon
Full edition · version 1.0

Ready to turn voices into text?

Download the complete Mac app. Add only the local models you want from its built-in model library.

Download TranscribeIt Local ↓English manual ↗Swedish manual ↗
DMG · 4.4 MBApple SiliconmacOS 14+