You join the call, the interviewer starts talking, and the assistant shows nothing. Or it catches one question and then goes quiet. This is the most common support problem for every desktop interview assistant, and it is rarely a broken app. A desktop assistant listens to the audio your PC plays. If the meeting sends that audio to a different device than the one being captured, the assistant hears silence.
The fixes below are ordered by how often they solve the problem. Most people are done after the first three. Work through them before a real interview, with the same headset, meeting app and network you will actually use. A five-minute test the day before is worth more than any setting you change during the call.
The steps use CrackInterviewAI on Windows 10 and 11 as the example, but the audio causes apply to any tool that captures system audio: Windows routing, Bluetooth profiles and meeting-app speaker settings behave the same way for all of them.
First, find out which half is broken
Open a YouTube video or any clip with clear speech and play it through the speakers or headset you plan to use. If the assistant transcribes the video, capture works and the problem is the meeting's audio route. Go to fixes 1 to 4. If it does not transcribe the video either, the problem is the Windows output device, volume or network. Go to fixes 5 to 9.
Also type one question instead of speaking it. In CrackInterviewAI, press CapsLock to open input mode, type the question and generate an answer. If the typed answer arrives, your account, balance and internet connection are fine and only audio capture needs attention. If the typed answer also fails, check sign-in and network before touching audio settings.
Fix 1: make the meeting play to your Windows default speaker
Meeting apps can choose their own speaker, separate from Windows. If Windows says your default output is "Speakers (Realtek)" but Zoom is set to your headset, the call audio never reaches the device being captured. Match them: pick the same device in the meeting app and in Windows Settings → System → Sound → Output.
Where to change it:
- Zoom: Settings → Audio → Speaker. Choose "Same as System" or the exact device Windows uses.
- Microsoft Teams: Settings → Devices → Speaker. Select the same device Windows shows as default.
- Google Meet: Settings → Audio → Speakers. Chrome and Edge usually follow the Windows default unless you picked a device here. Restart the test after each change.
Fix 2: the Bluetooth "Hands-Free" trap
Bluetooth headsets show up in Windows as two playback devices: a stereo "Headphones" entry and a "Headset" or "Hands-Free" entry that carries the microphone. When a call turns on the headset mic, Windows moves the call audio to the Hands-Free entry. If Windows still lists the stereo entry as default, a capture tool can end up listening to the idle device while the interviewer's voice goes to the other one.
Three reliable fixes: select the "Headset" (Hands-Free) entry as both the meeting speaker and the Windows default before the call; or keep the Bluetooth headset for listening only and use the laptop or a wired microphone for speaking, which keeps the headset in stereo mode; or use a wired headset for important interviews. Hands-Free mode also lowers audio quality, which can reduce transcription accuracy, so wired audio is the safest choice when it matters.
Fix 3: check the default communication device
Windows keeps two defaults: the default device and the default communication device. Some calling apps use the communication device, so the call can go somewhere other than the default output. Open the classic Sound control panel (Windows Settings → System → Sound → More sound settings), look at the Playback tab and make sure the same device carries both the "Default Device" and "Default Communications Device" labels. Right-click a device to set either one.
While you are there, remove confusion: disable playback devices you never use, such as an unused monitor speaker or an old virtual audio device. Fewer endpoints mean fewer places for call audio to disappear to.
Fix 4: switching headphones in the middle of a call
If you plug in or connect headphones after the call starts, Windows changes the default output and the meeting may or may not follow. CrackInterviewAI detects a new default output device and reconnects capture, but the meeting app has to be playing to that same device. After any switch, look at the meeting's speaker setting again and confirm the transcript is moving with a short "can you hear me?" exchange.
The safest habit is simple: connect the headset first, then join the call, then start the assistant. Do not change audio devices during an interview unless something is already broken.
Fix 5: volume, mute and the Volume Mixer
Capture tools receive the audio at the level Windows plays it. If the meeting app is muted or nearly silent in the Volume Mixer (Settings → System → Sound → Volume mixer), the assistant gets almost nothing to transcribe. Keep the meeting app and system volume at a normal listening level. Very quiet audio is one of the main reasons for missed words and half-captured questions.
Also check the physical controls: a headset volume wheel turned to zero, or a monitor's speakers muted, can leave Windows playing to a device that outputs nothing you can hear. If you cannot hear the interviewer yourself, fix that first.
Fix 6: exclusive mode, enhancements and virtual devices
Some apps take exclusive control of an audio device, which bypasses the shared Windows mix that capture tools listen to. In the Sound control panel, open the device's Properties → Advanced and untick "Allow applications to take exclusive control of this device". Then restart the meeting app.
Audio suites and virtual devices such as manufacturer "sound enhancement" apps, noise-removal tools and virtual cables can add playback endpoints and silently redirect the call. If a virtual device appears in the meeting's speaker list, choose your real headset or speakers instead, or temporarily disable the virtual device and test again.
Fix 7: when audio works but answers feel slow
Delay has several parts: the tool waits for the interviewer to finish speaking, turns speech into text, then streams the answer. A slow or unstable connection adds time to every step. Before an interview, run a speed test, pause large uploads, cloud backups and game downloads, and prefer a wired connection or strong Wi-Fi over a weak mobile hotspot.
Long, rambling questions also take longer to settle. In MANUAL mode you decide when the question is complete. Use AUTO only when your transcript is clean and you want the answer to begin as soon as transcription finishes. If an answer is late, a typed follow-up with the key words is often faster than waiting for audio.
Fix 8: keep a typed fallback ready
Every audio setup can fail on a bad day. Keep the typed path in muscle memory: open input mode, type the key words of the question, and generate. For coding prompts shown on screen, use screenshot mode instead of audio. Practising both fallbacks once is enough to stay calm if the audio route breaks mid-interview. The shortcut workflow guide covers the keys.
Fix 9: what to send support if nothing works
If a video transcribes but the meeting does not, or nothing transcribes at all, contact support with: the app version, Windows 10 or 11, your headset model and whether it is Bluetooth or wired, the meeting app, and the device names shown in Windows Sound settings and in the meeting's speaker list. That information usually pinpoints the route in one reply.
For setup basics, see the Windows setup guide and the Windows compatibility checklist. Always test with the free minutes on the exact PC you will use before buying time for a real interview.
Test your audio route before the real interview
New accounts include 30 free minutes. Run one video test, one practice call and one typed question on the PC you will use.
Frequently asked questions
Why does my interview assistant show no transcript?
Usually the meeting plays audio to a different device than the one being captured. Set the meeting app speaker and the Windows default output to the same device, then test with a video.
Does a Bluetooth headset work with an AI interview assistant?
Yes, but when the headset mic is on, Windows switches to a separate Hands-Free device. Select that same Headset entry as the meeting speaker and Windows default, or use a wired headset for important calls.
Why does the assistant hear YouTube but not Teams or Zoom?
The meeting app has its own speaker setting. Change it to the same device Windows uses as default output, or to Same as System in Zoom, and check the default communication device.
Why are answers slow even when transcription works?
Network speed, long questions and waiting for the speaker to finish all add delay. Use a stable connection, pause big downloads, and use MANUAL mode or a typed follow-up when you need control.
Can I use the assistant if audio capture fails during a call?
Yes. Type the question in input mode or capture an on-screen prompt with screenshot mode. Both work without audio capture.
Keep exploring
Return to the CrackInterviewAI homepage to download the Windows app, or browse all guides on the interview prep blog.