Use cases

Transcribe Teams or Zoom calls without a bot

You do not need a bot in the call to get a transcript. If you can hear the meeting on your PC, you can record what plays through your speakers or headset and transcribe it. No one sees a new participant, and you do not depend on the host's plan or settings.

Hand-drawn illustration of a video call with four participants and a speaker pill with a sound wave below, on a sage background

Why built-in transcripts often fail you

In Teams, transcription has to be allowed by your organisation's meeting policy, and the person who starts it needs the setting enabled too. Translated transcripts need Teams Premium.

In Zoom, the audio transcript comes from a cloud recording, which needs a paid plan and the Create audio transcript setting. Recordings saved to your computer do not get one.

As a guest in someone else's meeting, you rarely control any of this. Bot-based notetakers join the call as a visible participant, and some organisations block them.

Record what your PC plays

WhisperARC can capture the Windows output you choose, such as your headset or speakers. Open Settings > Audio capture, select that output and turn on Include my microphone so your own voice is in the transcript.

Start and stop the capture with its shortcut (Ctrl + W on a new installation) or from the WhisperARC bubble. When it stops, the transcript and the recording go to History.

  • Pick the device the meeting app actually uses, not just the Windows default.
  • Keep the volume at a normal level and do not switch outputs mid-call.
  • For long meetings, keep the laptop plugged in and awake.

Choose where the transcription runs

In Local Mode, Whisper runs on your PC and the meeting audio never leaves it. This is the safest choice for confidential calls.

Cloud sends the audio to WhisperARC's servers in Europe and is lighter on older PCs. With your own OpenAI key, the audio goes to OpenAI. The active glossary applies in every mode, so client names come out right.

Use the transcript after the call

Open History to search the transcript and replay any passage at up to 2× speed. Export DOCX for a document, TXT for notes or SRT and VTT when you need timestamps.

The transcript stays in the language spoken. Whisper does not label speakers; only some OpenAI models return speaker labels. Check names and decisions against the recording before you share a summary.

Sources & further reading