How to record a meeting on Mac and Windows

Recording your own voice is easy — every laptop has a microphone. Recording the other side of a call is where most people get stuck. They point a phone at the speakers, or they put the meeting on speakerphone and let the microphone pick up a thin, echoing version of everyone else.

You do not need to do either. Both macOS and Windows can hand an application a clean digital copy of the audio your computer is already playing. This guide explains how that works in plain terms, what to say before you press record, and the exact steps in JesRecap.

Two tracks, not one

A good local recording of a call is really two recordings taken at the same time:

Your microphone — everything you say, straight from your input device.
System audio — everything your computer plays out, which on a call means everyone else.

Keeping them separate is worth the small extra effort. Each track has its own volume, its own noise profile, and — importantly — its own identity. Because your voice arrives on a dedicated track, software can label your words as yours without guessing. JesRecap writes microphone.wav and system-audio.wav during the meeting, then mixes a third meeting.wav afterwards for comfortable playback. All three stay on your disk.

How macOS captures the other side: Core Audio process taps

For years, capturing system output on a Mac meant installing a virtual audio driver — a fake output device that quietly forwarded everything to a recorder. It worked, but it involved a kernel-level extension and it hijacked your audio routing.

Modern macOS offers something cleaner. Core Audio can create a process tap: think of it as a read-only listening port on the audio a specific application (or the whole system) is sending to your speakers. The audio continues to your headphones untouched; the tap simply receives a copy of the same samples. No driver to install, no re-routing, nothing to undo later.

macOS gates this behind an explicit permission — the same Audio Capture consent you would grant a screen recorder. You approve it once, in System Settings, and you can revoke it at any time. JesRecap needs macOS 14.2 or newer and runs natively on both Apple Silicon.

How Windows captures the other side: WASAPI loopback

Windows has had the equivalent capability built into its audio stack for a long time, under the name loopback capture. Normally an app opens an output device to play sound into it. In loopback mode, an app opens that same output device for reading and receives whatever mix is currently being sent to it.

The practical effect matches the Mac: a direct digital copy of the call audio, at full quality, with no "Stereo Mix" tinkering in the sound control panel and no virtual cable to install. JesRecap uses this on Windows 10 and 11.

Neither approach involves a bot. Nothing joins your Zoom, Teams, or Meet call. The recording happens one layer below the meeting app, on your own machine — which is why it works with any conferencing tool, and with in-person conversations through the microphone alone. More on bot-free recording →

Before you press record: consent

Recording law varies by country, state, and sometimes by industry, and this is not legal advice. But the professional habit is simple and it has never cost anyone a deal:

Say it out loud, at the start. "I'm recording this so I can write up notes afterwards — is that alright with everyone?" Ten seconds, on the record, before anything substantive is said.
Say what happens to it. People relax considerably when you can tell them the audio stays on your laptop and is not uploaded to a service. With a local recorder, that sentence is true.
Get a yes from everyone, not just the host. If one participant is uncomfortable, take written notes for that call instead.
Honour deletion requests. Know where the files live so you can actually remove them. In JesRecap, deleting a meeting from the library removes its audio and transcript from disk.
Check the rules for your context — some jurisdictions require all-party consent, and regulated fields (healthcare, legal, finance) often have their own retention policies. Ask once, follow it forever.

Recording a meeting in JesRecap, step by step

Install and grant permission. On first launch, JesRecap asks for microphone access and — on macOS — Audio Capture. There is a shortcut to the right settings panel in the app.
Pick a transcription model. The app downloads one Whisper-family model once, then reuses it offline forever. Start with the balanced option; you can switch later.
Choose your sources. In Audio sources, confirm your microphone and your speaker or headset. The mic check meter tells you whether you are actually being picked up before the meeting rather than after.
Press New meeting, then Start recording. Two live meters show both tracks, so you can see at a glance that the other side is arriving. Live transcript is optional — leave it off if you would rather keep memory use low and transcribe at the end.
Ask for consent, then run your meeting. Nothing appears in the participant list, so it is on you to mention it — see above.
Stop recording. JesRecap finalises both WAV files, mixes the playback track, transcribes locally, and — if speaker labels are enabled — separates "You" from the other voices with a local model.
Use the result. Copy the transcript, run a meeting action to paste a summary prompt into the AI you prefer, or export Markdown notes, plain text, a JSON archive, SRT/VTT captions, or the raw audio.

Small things that make recordings better

Wear headphones. Without them, the other side comes out of your speakers and back into your microphone, so both tracks contain everyone. Headphones keep the separation clean, which improves both playback and speaker labels.
Set your input before the call, not during it. Switching devices mid-recording is the most common cause of a half-silent track.
Do not worry about the mix. Because both source tracks are kept, you can retranscribe a meeting later — with a different model, for example — without re-recording anything.
Watch your disk, not your bandwidth. Local WAV is uncompressed and honest. Long recordings take space; nothing takes upload time.