You need two different kinds of capture: your microphone, and the system audio. Your microphone records you. It does not record the other people, because their voices are not in your room, they are audio your computer is playing through the speakers. Capturing that is called system audio or loopback capture.
If a recorder captures only the microphone, you end up with your own half of the conversation and silence where the answers were. This is the single most common way a meeting recording fails.
Recording a microphone is a normal thing every operating system supports. Recording what the computer is playing is a different, more awkward capability, and it works differently on Windows and Mac.
On Windows there is a supported way to capture the audio a device is playing, called WASAPI loopback. On Mac it is harder, which is why some tools are Windows-first on this feature, and why some Mac screen recorders cannot capture the other participants at all.
The result is that plenty of general-purpose screen recorders capture the screen and the microphone, and quietly leave out the half of the meeting you did not say.
Once you are capturing both, there is a choice: mix them into one audio track, or keep them apart. Keeping them apart is much better and it is worth insisting on.
Everything on the microphone track is you, by definition. Everything on the system track is everyone else, by definition. That gives perfect separation between you and them with no AI involved and no guessing. Software then only has to tell the other participants apart from each other, which is a much smaller problem than working out who is talking from a single mixed track.
Mixed into one track, all of that is lost, and the notes can no longer tell your commitments from theirs.
This is the part people skip. Test it once, properly, before a real call.
Better still, use a tool that shows you live. Jotted displays two meters while recording, one for the microphone and one for the system audio. If the system meter moves when the other person talks, their audio is being captured, and you know it during the call rather than after it. That feature exists because the founder once lost the other half of a client call and only found out afterwards.
If you use speakers rather than headphones, your microphone will also pick up the other person's voice coming out of them. It still works, but the two tracks bleed into each other and the separation gets messier. Headphones give you a clean separation and a better transcript.
Everything your computer is playing through its speakers or headphones. On a call, that is every other participant. It is a separate thing from your microphone and has to be captured deliberately.
Because the tool captured your microphone but not the system audio. This is the most common failure in DIY meeting recording setups.
Two. Keeping them separate gives you perfect separation between yourself and everyone else, which is how the notes know which commitments were yours. Mixing them throws that away.
Not required, but they help. On speakers, your microphone also picks up the other person's voice, so the two tracks bleed and the transcript labels get less clean.
The capability differs. Windows has a supported loopback capture path. Mac is harder, which is why some tools support it later there, and why some Mac screen recorders cannot capture other participants at all.
Record free forever, no card and no bot. Join the waitlist and you'll be first in.
Free forever plan · Unlimited recordings · No bot, ever