Features and settings

Voice questions and answers

Set your role and audio setup, then practice voice answers, spoken problem solving, scrolling and display translation.

Before practicing, learn the shortcuts and check basic settings.

Voice mode

Voice transcribes system audio and microphone input with automatic language detection. Available voice minutes depend on your plan. Complete permission and audio checks, then choose your role and audio setup before the interview.

Open Settings from the toolbar and choose My role: Candidate or Interviewer. Then open Voice Settings and choose Audio setup:

  • Computer call: the meeting or practice audio plays on your computer. Computer audio and microphone input are captured separately. Use this option for computer speakers too.
  • Phone speaker: a call plays through your phone speaker. Both voices enter the computer microphone and are labeled Mixed conversation; system audio is not captured.

Your role and audio setup save automatically. Wait for saving to finish, then click Voice on the toolbar to start recording. These two settings cannot change during recording; click Stop before changing them. The microphone and speaker controls mute or enable sources for the current session without changing speaker identity. A new recording enables sources according to the saved audio setup.

Voice settings

Before recording, open Voice Settings from the toolbar settings. With the voice panel open, use Voice Settings at the top right of the answer area to expand or collapse the settings section. Audio setup comes first, followed by microphone and computer audio selection and tests. You can also configure the answer model, response language, Knowledge Base, Concise answer, extra instructions, hotwords, display translation and scroll distance. My role is configured in the main toolbar settings.

SettingSelection and check
MicrophoneChoose your input device, or keep the default to follow the system. Click Test and speak while watching the level meter.
ComputerFor Computer call, choose the headphones or speakers that play the meeting, then play audio and test. Phone speaker does not use this source.

Device choices save automatically. Stop recording before changing devices. Use Refresh if a newly connected device is missing. Tests stop after about 10 seconds or when you click Stop. They only check local audio; confirm that transcription works once recording starts.

The row below the toolbar also provides quick controls for the answer model, response language, Knowledge Base and Concise answer. Changes save automatically; there is no separate Save button. Press Enter to add hotwords, using commas to separate multiple entries.

Response language applies to both voice Q&A and Voice solve. When code is involved, the selected Programming language is also provided. These preferences do not change automatic speech-language detection or Display translation.

With Concise answer off, questions use the normal answer mode. Turning it on uses concise answers. It applies only to Answer question, or Analyze follow-up for interviewers; it does not affect Voice solve or Assess solution.

Scroll distance is a percentage of the visible answer area. It defaults to half a screen (50%) and ranges from 10% to 100% in 10% steps. Reset restores 50%. Previous/Next buttons and their shortcuts use this distance, which adapts when you resize the window.

INTERACTIVE WALKTHROUGH

Try it: configure audio devices and voice answers

Click the numbered markers to see what happens next in the client, or use the tutorial controls below.

Current client · macOS interface · example values may be Chinese

01 / 11

Click the gear on the toolbar

Click the gear on the toolbar
1 → Click the settings gearFull screenshot (new tab)

Find the desktop client's floating toolbar and click the gear at its right edge.

What you should see: The settings panel opens with languages, models, capture mode and links near the bottom.

These steps use fixed examples. Click the yellow image marker or use Next tutorial step below.

Show all illustrated steps
  1. 1. Click the gear on the toolbar

    Find the desktop client's floating toolbar and click the gear at its right edge.

    Click the gear on the toolbar

    The settings panel opens with languages, models, capture mode and links near the bottom.

    These steps use fixed examples. Click the yellow image marker or use Next tutorial step below.

  2. 2. Open Voice Settings near the bottom

    Click Voice Settings at the bottom of the settings panel.

    Open Voice Settings near the bottom

    My role is in the main settings. Voice settings contains Audio setup, microphone and computer audio devices, local audio tests, Answer model and response language, followed by Knowledge Base, Concise answer, extra instructions, hotwords, display translation and scroll distance.

    Before recording, choose Candidate or Interviewer and Computer call or Phone speaker. Identity and audio setup cannot change during recording. Screenshot requirements live separately in Advanced Settings.

  3. 3. Choose the microphone you use

    Expand Microphone and choose your input device. This example selects a headset microphone; keep the default option to follow the system.

    Choose the microphone you use

    The choice saves automatically and is used for subsequent recordings.

    Use Refresh if a newly connected device is missing. Stop recording before changing devices.

  4. 4. Choose the output that plays the call

    Expand Computer and choose the headphones or speakers playing the meeting. This example selects headphones.

    Choose the output that plays the call

    Computer call captures the other person from the selected output.

    This does not change your meeting app’s output; both selections should match. Phone speaker only needs the microphone. If selection is unavailable, check OS support and permissions.

  5. 5. Test audio before recording

    Click Test beside Microphone and say a sentence. For Computer call, also play meeting or practice audio and test Computer separately.

    Test audio before recording

    A level meter appears and Test changes to Stop.

    This is a local capture check, not transcription. It stops after about 10 seconds or when you click Stop. This website uses simulated audio and does not access your microphone.

  6. 6. Check the level meter, then stop the test

    Watch the meter while speaking, then click Stop. Check Computer the same way.

    Check the level meter, then stop the test

    The test ends; continue configuring or start recording.

    If there is sound but no level, check device selection, mute settings and permissions, then test again. A moving meter confirms local audio only; check transcription after recording starts.

  7. 7. Enter a specific instruction for voice answers

    Find Extra instructions and enter a short instruction, such as giving the conclusion first and then a project example.

    Enter a specific instruction for voice answers

    The text remains in the field and is saved automatically.

    The sample instruction in the image is Chinese. No separate Save button is needed. Clear the field to remove the instruction.

  8. 8. Enter technical terms, then press Enter

    Type Redis,Kubernetes in the hotword field, then press Enter. Commas or semicolons can separate multiple terms.

    Enter technical terms, then press Enter

    The input clears and Redis and Kubernetes appear as tags below it.

    Text still sitting in the input has not been added. Use a tag's × to remove it. Check the counter for the current limit.

  9. 9. Enable Concise answer and review display settings

    Enable Concise answer below the model and response language. It affects Answer question, or Analyze follow-up as the interviewer, without changing Voice solve or Assess solution.

    Enable Concise answer and review display settings

    Concise answer is enabled; turn it off for ordinary answers. Your hotword tags remain saved.

    Display translation offers Off, Auto, ZH to EN and EN to ZH for Ultimate users; it does not replace response language. Adjust Scroll distance with − / + from 10% to 100% of the answer area, or reset to the default 50%.

  10. 10. Close Voice Settings, then reopen it

    Click × at the top right, return to the settings panel, then click Voice Settings again.

    Close Voice Settings, then reopen it

    The requirements and hotword tags remain after reopening.

    These voice fields save on change. Solution requirements in Advanced Settings use a separate Save action.

  11. 11. Test recognition and answer style separately

    Check the saved instruction and tags, close the window, and start desktop recording. Say a short sentence containing the terms, check the transcript, then trigger an answer using the voice walkthrough.

    Test recognition and answer style separately

    You can evaluate recognition accuracy and the requested answer style independently.

    Hotwords are recognition hints, not guarantees. Change one option at a time when troubleshooting.

Transcribe, generate and page through answers

Transcripts appear on the left and AI responses on the right. Once the question and its important conditions are transcribed, choose the appropriate action:

  • Answer question: handles general interview questions. For the Interviewer role, it becomes Analyze follow-up, which analyzes the candidate's answer and suggests a follow-up.
  • Voice solve: handles spoken programming, math, logic or scenario problems. For the Interviewer role, it becomes Assess solution, which gives a reference solution and points to verify. Code is generated only when the task requires a program, function, SQL query or code change.

Both actions use the current answer target and recent conversation background. The conversation record and recent transcript share a background budget of up to 2000 characters, keeping newer material first. Speaker labels, line breaks and truncation markers count toward that budget. The current answer target, reference material and prompt instructions are provided separately, so 2000 characters is not a limit on the full request or response.

INTERACTIVE WALKTHROUGH

Try it: voice questions and answers

Click the numbered markers to see what happens next in the client, or use the tutorial controls below.

Client 2.4 series · macOS interface · fixed practice dialogue in Chinese

01 / 10

Choose My role in the main settings

Choose My role in the main settings
1 → Choose CandidateFull screenshot (new tab)

Open toolbar Settings and choose Candidate or Interviewer. Then open Voice settings to configure audio capture.

What you should see: The role saves automatically and determines speaker labels and answer behavior for the session.

Complete this before recording. Interviewers should use the candidate’s resume as reference material.

Show all illustrated steps
  1. 1. Choose My role in the main settings

    Open toolbar Settings and choose Candidate or Interviewer. Then open Voice settings to configure audio capture.

    Choose My role in the main settings

    The role saves automatically and determines speaker labels and answer behavior for the session.

    Complete this before recording. Interviewers should use the candidate’s resume as reference material.

  2. 2. Choose the audio setup for this recording

    Open Voice settings and choose Audio setup. Use Computer call for meetings or audio played on your computer. Use Phone speaker when both voices from a phone call enter the computer microphone.

    Choose the audio setup for this recording

    The setup saves automatically. Computer call captures computer audio and microphone input separately; computer audio is the interviewer for candidates, or the candidate for interviewers. Phone speaker uses only the microphone and labels both voices Mixed conversation.

    Wait for saving to finish before recording. Role and audio setup cannot change during recording. This example uses Candidate and Computer call; interviewers see the corresponding analysis actions.

  3. 3. Check devices and levels before recording

    Choose microphone and computer audio devices in Voice Settings, then test each. Speak or play audio, check the meter, then stop the test.

    Check devices and levels before recording

    After the test ends, close Voice Settings and start recording.

    Phone speaker only needs the microphone. Tests check local audio, not transcription; this website uses simulated audio. The Voice Settings walkthrough above shows device selection.

  4. 4. Click Voice on the toolbar to open the voice panel

    Open your meeting or practice audio and complete the required microphone and screen-recording permissions. Once your audio setup has saved, click Voice on the toolbar.

    Click Voice on the toolbar to open the voice panel

    A timer, microphone and speaker controls, and Stop appear. The voice panel opens below and transcripts begin appearing on the left.

    This demo uses a fixed conversation and does not record you. Source controls apply to the current session; Phone speaker does not use system audio. New recordings enable sources according to the saved setup.

  5. 5. Select Manual for your first practice run

    Click Manual in the trigger controls below the toolbar. This example starts with Auto selected; after switching, you decide when to generate an answer.

    Select Manual for your first practice run

    Manual becomes selected. Transcription continues, and you decide when to generate an answer.

    Computer call and Phone speaker both support Manual and Auto. Auto processes the other person's new finalized transcripts in Computer call; finalized mixed transcripts in Phone speaker attempt to trigger when conditions are met. Choose Manual to wait for a complete problem. Voice solve is triggered manually.

  6. 6. Once the question appears, click Answer question

    Check that the question and its important conditions are transcribed on the left, then click Answer question at the top of the answer area. Use Voice solve for spoken programming, math, logic or scenario problems. Interviewers see Analyze follow-up and Assess solution instead.

    Once the question appears, click Answer question

    The answer panel starts generating a response.

    The buttons are disabled without a transcript. Selecting Manual alone does not generate an answer. Concise answer affects questions or follow-up analysis, not Voice solve. The default Voice solve shortcut is ⌘ / Ctrl + J.

  7. 7. Wait for the response

    No further client action is needed. The question has been submitted; wait for the answer to arrive.

    Wait for the response

    Text begins appearing in the answer panel.

    The conversation record and recent transcript share up to 2000 characters of background, keeping newer material first. The current answer target is provided separately. Timing depends on the model and network; avoid repeated triggers.

  8. 8. Click Next at the top right to read more

    After reading the visible answer, click Next in the voice answer panel. Use Previous to return to earlier text.

    Click Next at the top right to read more

    The answer display moves down to reveal more content. No new question is submitted.

    Scroll distance in Voice Settings defaults to half a screen (50%) and adjusts in 10% steps up to one screen. For another question, wait for its transcript and click Answer question. Default paging shortcuts use comma and period.

  9. 9. Click Stop to end this recording

    Save any content you still need to read, then click Stop on the toolbar.

    Click Stop to end this recording

    Once recording finishes, the timer stops, the voice panel closes and Voice returns to the toolbar.

    The next recording starts a new set of transcripts and answers. Use source controls for temporary muting. End recording before changing your role or audio setup.

  10. 10. Your voice practice is complete

    Check that Voice has returned to the toolbar. Before the next practice, check your saved role and audio setup, then start recording.

    Your voice practice is complete

    You have completed: configure and test audio → record → trigger manually → read the response → finish.

    Replay only restarts this website tutorial. Ultimate users can enable Display translation in Voice Settings; response language is configured separately. Continue to Knowledge Base management to use personal material.

Both Computer call and Phone speaker support Manual and Auto. These controls set when general answers are generated; transcription continues in either mode. In Computer call, computer audio captures the other person: the interviewer for the Candidate role, or the candidate for the Interviewer role. Auto processes new finalized transcripts from that target speaker. In Phone speaker, both voices share one microphone stream; new finalized mixed transcripts attempt to trigger a response when the trigger conditions are met. Choose Manual if you want to wait for a complete question or problem before generating a response. Voice solve is always triggered manually. Read the trigger rules and verification steps.

Real-time translation

Display translation (real-time translation) is available to Ultimate users. In Voice Settings, choose Off, Auto, ZH to EN, or EN to ZH. Translation is display-only: transcripts and AI answers keep their original text, and answer generation does not change.

Open Voice Settings → find Display translation → choose Off, Auto, ZH to EN or EN to ZH. If the options are unavailable, check Ultimate membership. Your choice saves automatically; inspect a transcript or completed response to see the result. If translation of a completed response fails, use Retry translation. Response language controls the language of AI answers and is configured separately.

See the Voice Settings screenshots.