Start with the right installer
Download the Apple Silicon build for an M-series Mac or the Intel build for an Intel Mac. Install Copycat, then open Settings before starting a session. The speech model and the model that generates answers are separate choices: getting one working does not automatically configure the other.
Capture the meeting’s sound
Copycat uses meeting/system audio as its default interviewer channel. The other participant’s voice must actually be playing through the meeting application. Your own microphone is a separate source: do not expect a system-audio-only session to transcribe speech that never reaches the computer’s output. A meeting app’s echo, playback, or monitoring settings can affect which voices appear.
Review macOS capture permissions
Allow the screen/system-audio capture permission requested by the installed app. The exact System Settings label depends on your macOS version. If you grant permission while Copycat is running, follow any macOS instruction to quit and reopen the app. A permission previously given to a different build or app identity may not apply to your current installation.
Install a speech model or configure a provider
In Settings → Audio, choose an installed local speech model or configure a supported cloud speech provider. Wait until a local model finishes downloading before selecting it. Use the provider’s supported language settings and valid credentials for cloud transcription. Choose based on a short test on your own Mac rather than a promised universal speed or accuracy ranking.
Test the live meter before relying on the transcript
Start a session and play a short piece of speech from the meeting application. Look at Copycat’s real audio-level meter. If the meter stays flat, investigate capture permissions, playback, and device routing first. If it moves but no words appear, inspect the selected speech provider, installed model, language, and any provider error.
Use Listen to retry
Use the overlay’s Listen control or its configured shortcut to retry interviewer listening after fixing a setup issue. Shortcut combinations can be customized, so check Settings instead of assuming that every installation uses the same keys. Leave time to test before an important session.
If you see “waiting for interviewer audio”
Check that the other participant is speaking and that the meeting app is playing their audio. Confirm capture permission for the installed Copycat app, then verify the speech model is ready. Retry Listen after correcting the problem. For cloud providers, check connectivity, credentials, quota, and provider errors. For local models, try a smaller supported model if the device is struggling.
Keep capture expectations clear
A live meter shows received audio; an example waveform on a marketing page does not prove capture works on your device. Use the app’s actual meter and transcript to test. Get consent where required before recording or transcribing other people, and review generated text before relying on it.