KONWA setup

# KONWA setup

KONWA translates what the user says, speaks the translation, and sends it through a virtual audio device into another app, so the people in a call, a stream or a game hear the translation.

## The chain

Four hops, and all four are needed:

1. Real microphone → KONWA
2. KONWA → virtual audio device (installs with KONWA; nothing to download separately)
3. Virtual audio device → the target app’s microphone input
4. Target app → the other people

One path runs beside it: the target app’s output → the user’s headphones. That one must be headphones, not the virtual device.

Every fault is one broken hop. Find the hop first, then fix it.

## Steps

Find the operating system, the installed apps and the device list for yourself.

### 1. Install
Take the build for this system from konwa.live/download and install it.
Done when: KONWA launches.

### 2. Microphone permission
Confirm the system has granted KONWA microphone access.
Done when: KONWA no longer reports a missing permission.

### 3. Sign in
Email plus a six-digit code. There is no password.
Done when: KONWA shows a signed-in account.

### 4. Virtual device — hop 2
Open Settings → Devices, confirm the virtual microphone is listed, and read its device name.
Done when: you hold the exact device name. Later steps need it.

If it is not listed, tell the two causes apart before acting:
- The system blocked the install — check System Settings → Privacy & Security on macOS, Device Manager on Windows.
- Windows has not restarted since the install — restart and check again.

### 5. Input device — hop 1
In the same panel, set the user’s real microphone as KONWA’s input.
Done when: the panel shows the real microphone as the input, not the virtual device.

### 6. Target app — hop 3
Work out which app the user wants this in: look at what is installed or running, and offer them the candidates to pick from.
Set that app’s microphone to the device name from step 4, and confirm its output is still headphones.
Done when: both are checked — microphone is the virtual device, output is headphones.

Where the setting usually is:
- Discord: User Settings → Voice & Video
- Steam: Steam → Settings → Voice
- Zoom, Tencent Meeting: Settings → Audio
- Microsoft Teams: Settings → Devices
- OBS: add an Audio Input Capture source, and mute the existing real microphone source
- TikTok Live Studio: audio settings

### 7. Verify
Start translating in KONWA and have the user say a line.
Done when: the user confirms the translation went out, not their own voice.

## Finding the broken hop

| Symptom | Broken at |
| --- | --- |
| They hear nothing | Hop 3 is not connected, or KONWA is not translating |
| They hear the user’s own voice | Hop 3 still points at the real microphone |
| They hear the user twice, or an echo | The target app’s output is set to the virtual device |
| The level swells and fades | The target app is adjusting microphone level automatically; turn it off |
| The user hears nothing at all | The system output is set to the virtual device; set it back to headphones |
| KONWA receives no speech | Hop 1 is not connected, or the system has not granted permission |
| KONWA translates the other person too | The user is on speakers and the microphone picks them up; move to headphones |

## Boundaries

- KONWA translates only what the user says. Not other people, not game sound, not music.
- The translated voice arrives after the user’s own, usually within a second.
- Hop 2 is silent while KONWA is not translating: the target app still points at the virtual device, so no one hears the user. Set the microphone back to the real device when not translating.