Wake Word

The typical workflow for interacting with a voice assistant is to first activate it with a "wake" or "hot" word, then provide your voice command. Rhasspy supports listening for a wake word with one of several systems.

You can also wake Rhasspy up using the HTTP API by POST-ing to /api/listen-for-command. Rhasspy will immediately wake up and start listening for a voice command.

The following table summarizes the key characteristics of each wake word system:

System Performance Training to Customize Online Sign Up
pocketsphinx poor no no
precise moderate yes, offline no
snowboy good yes, online yes
porcupine excellent yes, offline no

Pocketsphinx

Listens for a keyphrase using pocketsphinx. This is the most flexible wake system, but has the worst performance in terms of false positives/negatives.

Add to your profile:

"wake": {
  "system": "pocketsphinx",
  "pocketsphinx": {
    "keyphrase": "okay rhasspy",
    "threshold": 1e-30,
    "chunk_size": 960
  }
},

"rhasspy": {
  "listen_on_start": true
}

Set wake.pocketsphinx.keyphrase to whatever you like, though 3-4 syllables is recommended. Make sure to train and restart Rhasspy whenever you change the keyphrase.

The wake.pocketsphinx.threshold should be in the range 1e-50 to 1e-5. The smaller the number, the less like the keyphrase is to be observed. At least one person has written a script to automatically tune the threshold.

See rhasspy.wake.PocketsphinxWakeListener for details.

Mycroft Precise

Listens for a wake word with Mycroft Precise. It requires training up front, but can be done completely offline!

Add to your profile:

"wake": {
  "system": "precise",
  "precise": {
    "model": "model-name-in-profile.pb",
    "sensitivity": 0.5,
    "trigger_level": 3,
    "chunk_size": 2048
  }
},

"rhasspy": {
  "listen_on_start": true
}

Follow the instructions from Mycroft AI to train your own wake word model. When you're finished, place both the .pb and .pb.params files in your profile directory, and set wake.precise.model to the name of the .pb file.

See rhasspy.wake.PreciseWakeListener for details.

Snowboy

Listens for a wake word with snowboy. This system has the good performance out of the box, but requires an online service to train.

Add to your profile:

"wake": {
  "system": "snowboy",
  "hermes": {
    "wakeword_id": "default"
  },
  "snowboy": {
    "model": "model-name-in-profile.(u|p)mdl",
    "audio_gain": 1,
    "sensitivity": 0.5,
    "chunk_size": 960
  }
},

"rhasspy": {
  "listen_on_start": true
}

Visit the snowboy website to train your own wake word model (requires linking to a GitHub/Google/Facebook account). This personal model with end with .pmdl, and should go in your profile directory. Then, set wake.snowboy.model to the name of that file.

You also have the option of using a pre-train universal model (.umdl) from Kitt.AI. I've received errors using anything but snowboy.umdl, but YMMV.

See rhasspy.wake.SnowboyWakeListener for details.

Porcupine

Listens for a wake word with porcupine. This system has the best performance out of the box. If you want a custom wake word, however, you will need to re-run their optimizer tool every 30 days.

Add to your profile:

"wake": {
  "system": "porcupine",
  "porcupine": {
    "library_path": "porcupine/libpv_porcupine.so",
    "model_path": "porcupine/porcupine_params.pv",
    "keyword_path": "porcupine/porcupine.ppn",
    "sensitivity": 0.5
  }
},

"rhasspy": {
  "listen_on_start": true
}

There are a lot of keyword files available for download. Use the linux platform if you're on desktop/laptop (amd64) and the raspberrypi platform if you're using a Raspberry Pi (armhf/aarch64). The .ppn files should go in the porcupine directory inside your profile (referenced by keyword_path).

If you want to create a custom wake word, you will need to run the Porcupine Optimizer. NOTE: the generated keyword file is only valid for 30 days, though you can always just re-run the optimizer.

See rhasspy.wake.PorcupineWakeListener for details.

MQTT/Hermes

Subscribes to the hermes/hotword/<WAKEWORD_ID>/detected topic, and wakes Rhasspy up when a message is received (Hermes protocol). This allows Rhasspy to use the wake word functionality in Snips.AI.

Add to your profile:

"wake": {
  "system": "hermes",
  "hermes": {
    "wakeword_id": "default"
  }
},


"rhasspy": {
  "listen_on_start": true
},

"mqtt": {
  "enabled": true,
  "host": "localhost",
  "username": "",
  "port": 1883,
  "password": "",
  "site_id": "default"
}

Adjust the mqtt configuration to connect to your MQTT broker. Set mqtt.site_id to match your Snips.AI siteId and wake.hermes.wakeword_id to match your Snips.AI wakewordId.

See rhasspy.wake.HermesWakeListener for details.

Command

Calls a custom external program to listen for a wake word, only waking up Rhasspy when the program exits.

Add to your profile:

"wake": {
  "system": "command",
  "command": {
    "program": "/path/to/program",
    "arguments": []
  }
},

"rhasspy": {
  "listen_on_start": true
}

When Rhasspy starts, your program will be called with the given arguments. Once your program detects the wake word, it should print it to standard out and exit. Rhasspy will call your program again when it goes back to sleep. If the empty string is printed, Rhasspy will not wake up and your program will be called again.

The following environment variables are available to your program:

  • $RHASSPY_BASE_DIR - path to the directory where Rhasspy is running from
  • $RHASSPY_PROFILE - name of the current profile (e.g., "en")
  • $RHASSPY_PROFILE_DIR - directory of the current profile (where profile.json is)

See sleep.sh for an example program.

See rhasspy.wake.CommandWakeListener for details.

Dummy

Disables wake word functionality.

Add to your profile:

"wake": {
  "system": "dummy"
}

See rhasspy.wake.DummyWakeListener for details.