Private Zoom Meeting Transcription Without the Cloud

The private way to transcribe a Zoom call is to use Zoom’s computer (local) recording, which saves the files to your own machine. Then transcribe the audio with a Whisper-based tool that runs offline, either on the same laptop or on your phone. Zoom’s cloud transcripts, its AI Companion summaries and third-party notetaker bots all process the conversation on someone else’s servers. That’s fine for a routine standup and a problem for a client negotiation.

This guide walks through the local route, the tools that do the transcription, and the settings that quietly undo your privacy if you miss them.

Where do Zoom’s built-in options send your audio? #

Bots also join as a visible participant. Otter’s free Basic plan, for example, includes 300 transcription minutes a month with a 30-minute cap per conversation, and its paid Pro plan is $16.99 a month or $8.33 a month billed annually, according to Otter’s pricing page. Whatever the price, the recording lives in Otter’s cloud.

Step 1: Record the Zoom meeting to your computer #

Zoom calls this computer recording. According to Zoom’s support page, the saved files can include video, shared content, audio files, chat messages and closed captions. Hosts can record, and participants can too if the account settings allow it.

1. Open the Zoom desktop app on Windows, Mac or Linux. Computer recording isn’t available on iOS, iPad or Android.
2. In the meeting, click Record, then Record on this computer. If you’re not the host, ask the host for recording permission.
3. When the meeting ends, Zoom converts the recording. Wait for it to finish before closing the app.
4. Find the files in the Zoom folder inside your Documents folder (the default location). You can change it in Zoom’s settings under Recording.

Two settings in the desktop app’s Recording section are worth checking before your next call:

• A separate audio file for each participant. If your account allows it, this gives you one track per speaker, which makes “who said what” trivial later.
• Save location. Keep recordings out of any folder that syncs to the cloud. More on that below.

Step 2: Transcribe the recording on your computer #

These tools all run Whisper locally. None needs an internet connection after the model download.

MacWhisper’s free version already transcribes M4A and MP4, recognizes speakers, and exports .txt, .srt and .vtt. Buzz does the same on Windows and Linux. Superwhisper can also send text to cloud AI models if you connect an API key, so stick to its local models for confidential calls.

If you prefer the command line, convert the Zoom audio to 16 kHz mono and run whisper.cpp:

ffmpeg -i audio_recording.m4a -ar 16000 -ac 1 -c:a pcm_s16le meeting.wav
./build/bin/whisper-cli -m models/ggml-large-v3-turbo.bin -f meeting.wav -otxt -osrt

The .srt output has a timestamp on every line, which makes it easy to jump back to the audio when a sentence looks wrong. Mac users will find a fuller walkthrough in how to transcribe meetings privately on a MacBook. There’s a tool-by-tool comparison in private audio transcription software for Mac.

Or transcribe the Zoom recording on your phone #

If you’d rather not install anything on a work laptop, move the audio file to your phone and transcribe it there. Private Transcribe imports audio and video files (m4a, mp3, wav, aac, mp4, mov) and runs Whisper on the phone’s own processor, so the recording never touches a server. It has no account, and after a one-time model download it works in airplane mode.

1. Copy the audio file from the Zoom folder to your iPhone or Android phone by cable, AirDrop or a local file transfer. Don’t bounce it through email or a cloud drive.
2. In the app, tap Import audio on the Record tab and pick the file.
3. Choose the Professional or Best model for an important call. They’re slower but handle crosstalk and accents better.
4. Open the transcript in History to read it with timestamps, then copy or share it. The one-time Pro purchase adds .txt and .srt export.

Imports are limited to 90 minutes per file. For a longer call, split it on the computer first:

ffmpeg -i audio_recording.m4a -f segment -segment_time 3600 -c copy part%02d.m4a

The app doesn’t label speakers. If you need names, use the per-participant audio files or a desktop tool with speaker recognition.

Can you get live, private captions during a Zoom call? #

Not easily. Zoom’s live captions and AI features run on Zoom’s side. For local live text you’d route Zoom’s audio through a virtual audio device and run a streaming engine such as whisper.cpp’s whisper-stream. We cover that capture setup in Google Meet transcription without bots, and it works the same for Zoom. Live local captions lag a few seconds and make more mistakes than a pass over the finished recording. The realistic trade-offs are covered in transcription accuracy for confidential business calls.

Getting speaker labels without the cloud #

You have three local options, from easiest to hardest:

1. Record a separate audio file per participant in Zoom, then transcribe each file. Every line is already attributed.
2. Use a desktop app with speaker recognition, such as MacWhisper or Buzz.
3. Run pyannote.audio on the mixed recording. Its community pipeline runs locally after a one-time model download with a Hugging Face token.

For two or three people who take turns, any of these works well. For a ten-person call where people talk over each other, expect to correct labels by hand.

Don’t let a sync folder undo your privacy #

This is the mistake that catches careful people. Zoom saves to Documents by default, and many computers sync Documents to the cloud without anyone thinking about it:

• Mac: iCloud Drive’s “Desktop & Documents Folders” option uploads everything in Documents.
• Windows: OneDrive folder backup often covers Documents.
• Work laptops: IT may back up the whole user folder.

Point Zoom’s recording location at a folder that doesn’t sync. Once the transcript is checked, delete the audio. A text file is small and easy to protect, while the recording holds everyone’s voice.

Consent and the recording notice #

Zoom tells participants when a meeting is being recorded, but that doesn’t settle the legal question. Consent rules vary by country and US state, and some require every participant’s agreement. Say at the start that you’re recording locally for notes and that nothing will be uploaded. Clients usually appreciate hearing it.

Frequently asked questions #

Is Zoom’s own transcription private? #

Zoom’s transcripts and AI Companion features run on Zoom’s cloud, so the audio and text are stored and processed by Zoom under your account’s settings. That’s acceptable for many companies. It isn’t “private” in the sense of never leaving your control. For that you need a local recording and local transcription.

Can I record a Zoom meeting on my phone and transcribe it? #

The Zoom mobile app can’t save a recording to the phone. You can record on a computer and move the file to your phone, or, for an in-person meeting, record the room with a phone transcription app.

How long does local transcription of a one-hour Zoom call take? #

It depends on the model and hardware. On a recent Apple silicon Mac or a PC with an NVIDIA GPU, a large Whisper model usually finishes well under the meeting’s length. On an older CPU-only laptop, a mid-size model can take about as long as the meeting. Smaller models are faster but less accurate.

Do other participants see that I’m transcribing? #

They see Zoom’s recording indicator when you record to your computer. The transcription step afterwards is invisible to them, because it happens entirely on your own device.