Voice Commands / Voice Controls

Beta: New Voice Command Engine

Starting with Yatse 8.3.5 a new powerful AI base Voice Command engine is in place. It currently support English and French but can be improved to support more languages:

See AI Voice Commands for more details

How to use

Voice commands can be started in a lot of way:

  • In the application : Press the microphone icon in the left menu (near to Remote button)
  • You can add a Voice Commands widgets to your home screen for direct action
  • If you are using Jelly Bean you can set Voice Commands to replace the Google now shortcut (swipe up of the home virtual Android button).
  • You can also use the api to start it from other apps (See Api - Use Yatse from other applications (Ex: Tasker))

Yatse voice commands currently support English / French / Portuguese / Italian / German / Polish / Russian and Dutch language, see the end of this page on how you can contribute to add support for your language in a future release.

Yatse tries to analyse what you say in a smart way to allow natural commands. Google Voice recognition engine is used to transform what you say in a list of possible sentences that Yatse then interpret to commands.

Due to limitations in Google Engine, when you use natural sentences in your native language to try to play English movie title or artists you may have to force the accent so Google do not try to change the title to some words in your language. There's nothing we can do about this.

If you find some command missing (related to those already implemented) please contact me with all the necessary information to add support for it (All grammar / gender formulation and logs see last part of this page)

English voice commands

Basic playback commands:

  • play -> play
  • pause -> pause
  • stop / end / finish -> stop
  • mute / unmute -> toggle mute
  • next -> next
  • previous -> previous
  • forward / faster -> forward
  • rewind / backward -> rewind
  • wake on lan / start kodi / start computer -> Wake on Lan

Natural language commands:

You can start your sentences with "I want to" this will be ignored in the parsing.

Remark: Some commands have default action for example you can say "listen to Massive Attack" to play all songs from this artist.
But for artists that are not well known you may have better result by saying "listen to the artist Massive Attack" so that Google engine know you're talking about an artist and send back better proposals.

The command listed here are not exhaustive you may want to try others, check the samples for ideas.

  • search -> Start global search on your terms with a filter on type if subcommand is (song / album / artist / movie / show / episode)
    Samples : "search for the artist Massive Attack", "search a movie named Avatar", ...
  • sync (or synchronize) -> Start Yatse media synchronization of specific type (can also use database as type for a global sync)
    Samples : "sync my movies", "I want to synchronize my episodes", "synchronize database",...
  • watch -> Start a movie named search words (if multiple movies are found, start the global search to list them)
    Samples : "I want to watch Avatar" , "watch Avatar"
  • watch movie -> Start a movie named search words (if multiple movies are found, start the global search to list them)
    Samples : "Watch the movie named Die Hard", "I want to watch the movie Avatar"
  • watch random movie (or just watch movie) -> Start a random movie
    Samples : "Watch a random movie", "I want to watch a movie"
  • watch -> Start the first unseen episode from show named search words
    Samples : "Watch the next Dexter", "I want to watch the next episode of Battlestar Galactica"
  • watch -> Start the last episode from show named search words
    Samples :"watch the latest episode of Battlestar Galactica", "watch the last Dexter"
  • listen -> Start playing specific type search words, will show global search on multiple match
    Samples : "Listen to Massive Attack" , "Listen to the album named Mezzanine", "I want to listen to the album Death Magnetic"
  • listen -> Start playing specific type search words with music as alias for genre for better detection. If no search words will play 50 random songs
    Samples : "I want to listen to some Rock", "Listen rock music" , "Listen to Mezzanine album"
  • listen random -> Will start a random play of selected type
    Samples :"listen to a random artist", "listen some music", "I want to listen to a random genre", ....

Remark : Watch and Listen can be replaced by more generic play or start command.

Other languages voice commands

All commands are the same as in English but in your natural language and with correct word ordering.
Since we can not write all possible sentences in all languages, we'll only display the supported words for each kind, for the full sentences you should know them better than us since it's your native language :)

Those lists are not exhaustive and does not show all gender / conjugations.


play : jouer , continuer
pause : pause
stop : stop, fin , arrêter
toggle mute : silence
next : suivant
previous : précédent
forward : avancer
rewind : reculer
search : chercher, rechercher
synchronize : synchroniser
watch : regarder
listen : ecouter
play : jouer
random : aléatoire
movie : film
show : série
episode : épisode
song : chanson
album : album
artist : artiste
genre : genre
music : musique
next : prochain
latest : dernier


play : riprendi , riproduci
pause : interrompi , pausa
stop : stop, ferma , termina
toggle mute : silenzio
next : prossimo , successivo
previous : precedente
forward : avanti
rewind : indietro
search : *cerca, ricerca*
synchronize : sincronizzar
watch : guarda
listen : ascolta
play : riproduci
random : casuale
movie : film
show : serie
episode : episodi
song : canzone
album : album
artist : artista
genre : genere
music : musica
next : prossimo
latest : ultimo


play : tocar
pause : pausar
stop : parar, terminar , finalizar
toggle mute : mudo , silenciar
next : próximo
previous : anterior
forward : avançar
rewind : retroceder , rebobinar
search : buscar
synchronize : sincronizar
watch : assistir
listen : ouvir
play : tocar
random : aleatório
movie : filme
show : programa
episode : episódio
song : faixa
album : album
artist : artista
genre : gênero
music : música
next : próximo
latest : último


play : speel
pause : pauze
stop : stop, ende , finish
toggle mute : geluid uit , geluid aan
next : volgende
previous : vorige
forward : vooruit spoelen
rewind : terug spoelen
search : zoek
synchronize : synchroniseer
watch : kijk
listen : luister
play : speel
random : random
movie : film
show : show
episode : aflevering
song : lied
album : album
artist : artiest
genre : genre
music : muziek
next : volgende
latest : laatste


play : spiele
pause : pause
stop : stop
toggle mute : ton an, ton aus
next : nächste
previous : früher
forward : vorspulen
rewind : zurückspulen
search : suche
synchronize : synchronisiere
watch : schaue(n)
listen : hören
play : spiele
random : zufälligen
movie : film
show : serie
episode : episode
song : lied
album : album
artist : künstler
genre : genre
music : musik
next : neueste
latest : letzte(n), nächste(n)


play : graj
pause : pauza
stop : stop
toggle mute : wycicz
next : następna, następny
previous : poprzedni, poprzednia
forward : przewiń, do przodu, naprzód
rewind : do tyłu, cofnij
search : szukaj, przeszukaj, wyszukaj
synchronize : synchronizuj
watch : oglądaj)
listen : słuchaj
play : włącz, graj
random : losowy
movie : film
show : serial
episode : część, epizod
song : piosenka, piosenkę
album : album
artist : artysta, artystę
genre : gatunek
music : muzyka, muzykę
next : następny
latest : ostatni

How to contribute (adding support for a new language or improving an existing one).

The goal of Yatse Voice Command is to be fun and natural to use. To achieve that it needs to parse natural language sentences.
To achieve natural commands Yatse does complex tree analysis and grammar / wording changes to be compatible with most language and usage.
To build the analysis tree, the grammar parser and the dictionaries we need data, lot's of data from real users saying real sentences.
The data needed can be split in two categories :

  • Natural language structure
  • Google voice analysis results

Natural language structure

For this part we need a document containing all possible sentences you think of for the supported command.

Example : (In the document you send us the bold part needs to be in your native language)

Listen some music - To start random music
Listen to some rock music - To start playing music of genre rock
Watch a movie - To start playing a random movie
Watch a random movie - To start playing a random movie
Watch the movie avatar - To start playing the avatar movie
Listen to the album white - To start playing the white album
Listen to artist massive attack - To start playing songs from massive attack
Listen the song xxxx - To start playing the song xxx
Watch the last episode of battlestar galactica - To start playing the next episode from battlestar galactica
Watch next episode of battlestar galactica - To start playing the next episode from battlestar galactica
Start the movie Avatar - To start playing the avatar movie
Play the movie Avatar - To start playing the avatar movie
Synchronize movies - To synchronize movies

Those commands are just a small list, the more commands you gives the better the engine will be.

(We also need formulations like I want to)

Beside that we need a dictionary of all basic commands and words (You can have multiple words for the same purpose like for example song and track for song)

Example of words needing a translation:
play, pause, stop, mute, unmute, next, previous, forward, rewind, movie, show, episode, song, album, artist, database, genre, music , next , last

And to finish we also need a list of articles or small words often used in sentences but that won't have any real impact on the commands or can cause problem in search.

For example words like the, a, an, to, of, and, I, it,....

Logs from Google Engine

For the second part of data we need logs from Yatse (Activate debug mode / see Wiki)
Since Google voice recognition is good but not perfect we need lots (really lots) of logs.
Simply start Yatse and use the voice command and say all the possible commands you think of with real media names.

Then send us the logs (You can contact us from the Help / About screen in Yatse ) all the logs will be treated anonymously and deleted after voice analysis is done.

This data is really important to match what Google sends back and what you say.

Go to top