Voice Control

Voice Control

EmoTracker includes a built-in voice control extension that lets you track items and manage locations entirely hands-free using offline speech recognition powered by Vosk.

Prerequisites

Voice control requires the PortAudio audio library to be present on your system. EmoTracker does not bundle it.

macOS

Install PortAudio via Homebrew:

brew install portaudio

Windows

Download the PortAudio binaries from the PortAudio website and place portaudio_x64.dll (or portaudio_x86.dll for 32-bit) in the same folder as EmoTracker.exe, or in a directory on your system PATH.

Linux

Install PortAudio from your distribution's package manager:

# Debian / Ubuntu
sudo apt install libportaudio2

# Arch Linux
sudo pacman -S portaudio

# Fedora
sudo dnf install portaudio

---

Enabling voice control

1. Open Settings (the gear icon in the status bar).
2. Under the Extensions section, enable Voice Control.
3. A microphone icon will appear in the status bar.

---

First run — downloading the speech model

Voice recognition requires an offline speech model (~40 MB). EmoTracker will prompt you to download it automatically the first time you activate the extension.

1. Click the microphone icon in the status bar and choose Active from the context menu, or enable it from Settings.
2. If no model is found, a download prompt appears. Click Download to begin.
3. A progress bar shows the download status. Once complete, the model is installed automatically and recognition starts.

The model is saved to your EmoTracker data directory (vosk-model subfolder) and only needs to be downloaded once.

To download it manually, grab vosk-model-small-en-us-0.15.zip from alphacephei.com/vosk/models, extract it, and rename the extracted folder to vosk-model inside your EmoTracker data directory.

---

Selecting an input device

Right-click the microphone icon in the status bar to open the context menu. The Input Device submenu lists all available audio input devices. Click a device to select it — your preference is saved and restored on next launch.

The system default input device is selected automatically if no preference has been saved.

---

Using voice control

Voice control uses a wake phrase before every command. Speak clearly at a normal pace.

Wake phrases

Wake phrase	Effect
hey tracker	Issues the following command
hey babe	Issues the following command (and earns a comment)

Item commands

Command	Effect
hey tracker track [item]	Toggle item on / advance progressive item
hey tracker track the [item]	Same as above
hey tracker toggle [item]	Toggle item on
hey tracker toggle [item] off	Toggle item off
hey tracker track [item] down	Downgrade a progressive item one step
hey tracker set [item] as [stage]	Advance a progressive item to a named stage
hey tracker add a [item]	Increment a consumable
hey tracker remove a [item]	Decrement a consumable

Location commands

Command	Effect
hey tracker clear [location]	Full-clear all available checks at a location
hey tracker reset [location]	Reset a location's check count
hey tracker pin [location]	Pin a location
hey tracker remove pin for [location]	Unpin a location
hey tracker mark [item] at [location]	Capture an item at a location
hey tracker mark [item] on [location]	Same as above

Options

Command	Effect
hey tracker enable show all locations	Show all locations regardless of accessibility
hey tracker disable show all locations	Restore normal accessibility filtering
hey tracker enable chat hud	Connect the Twitch chat HUD
hey tracker disable chat hud	Disconnect the Twitch chat HUD

Control

Command	Effect
hey tracker stop listening	Deactivate voice control
hey tracker undo that	Undo the last tracked change

---

How item names are matched

Item and location names are registered from the loaded pack. EmoTracker automatically attempts to split compound words that are not in the speech model's vocabulary — for example, hookshot is split into hook shot so the recogniser can hear it correctly.

Pack authors can also provide explicit alternate phrases using the phonetic_substitutes item property. Any substitute phrase triggers the same action as the item's canonical name.

---

Troubleshooting

The microphone icon doesn't appear : Make sure Voice Control is enabled in Settings.

Activating the extension shows no download prompt and does nothing : PortAudio could not be loaded. Verify you have installed it per the prerequisites above and that the library is on your system's library search path.

Commands are not recognised : Speak the wake phrase (hey tracker) clearly before the command. Background noise or a low-quality microphone can reduce accuracy. Try selecting a different input device from the microphone context menu.

The download fails or loops : Check that EmoTracker has write access to its data directory. If a partial vosk-model-download.zip is present there, delete it and try again.

A specific item name is never recognised : The item name may contain words not in the speech model's vocabulary. Ask the pack author to add a phonetic_substitutes entry for that item, or check the Authoring Items — Voice control page if you are the pack author.