Installing/upgrading the library
You get an error message similar to
E: Unable to locate package python3-venv
while trying to configure a new Python virtual environment.
On older versions of Ubuntu, the meta package for the Python 3 virtual environment (
python3-venv) is not available. Use a fully-qualified version instead:
sudo apt-get install python3-dev python3.4-venv
You get an error message similar to
No matching distribution found
<wheel> is not a supported wheel on this platform.
Check the device architecture:
Linux raspberrypi 4.9.24-v7+ #993 SMP Wed Apr 26 18:01:23 BST 2017 armv7l GNU/Linux
The library is only supported on devices with linux-armv7l and linux-x86_64 architectures. For other devices, use the Google Assistant Service instead.
When trying to generate credentials with the authorization tool, the tool prompts you to visit a URL and sign in to your Google account. The page then displays an error message (for example, "Something went wrong").
- Make sure you enabled the Google Assistant API and configured the OAuth consent screen in Configure a Developer Project and Account Settings.
The Google Assistant responds Actually, there are some basic settings that need your permission first...
- Set activity controls for the account that you used when setting the authorization code.
Voice recognition issues
Assistant responds I don't know how to help with that or I don't understand.
- Verify that your voice was recognized properly. Check the sample code output for a transcript of the user request. You can also look at my activity. You can play the audio received by the Assistant to check for audio quality issues. If you hear excessive noise or if the audio is distorted, check that your microphone is configured properly or try a different microphone.
Assistant does not recognize a trait that you added to a model.
Make sure the trait was added to the model (see Register Traits for the Device).
If you generated a device instance for a model, the device instance will not automatically update when its model is changed by other users. You need to manually re-register the device instance using the registration command line tool or the REST API. You can find the device instance ID in the output of the sample.
Device Actions issues
The Google Assistant responds saying that your app is not supported on your device.
- In the Actions Console, you can set the surface capabilities of your Assistant app. If you set screen output to be required, your custom actions will not function. The Google Assistant SDK does not support screen capabilities at this time.
Assistant does not seem to respond to sentences that begin with Ok Google or Hey Google.
If you still get little to no recognition of your voice, try calibrating your microphone sensitivity:
- Terminate (Ctrl+C) any sample code that is currently running.
- Run the following command:
- Run the sample code and try again. Go back to step 1, but keep lowering
ASSISTANT_MIC_SENSITIVITYvalue in intervals of -6 (e.g., -6, -12, -18, -24...) until you find that voice recognition is optimal. If you get to -60, check your microphone.
Alarms and timers issues
The Assistant does not follow verbal commands (for example, pausing or stopping a timer).
- This is an issue if you have the StartStop or OnOff trait enabled for a given device model. Remove this trait from the device model and verbal commands will work.
Podcast and news playback issues
The Assistant does not follow playback commands (for example, pausing or stopping a podcast).
- This is an issue if you have the StartStop or OnOff trait enabled for a given device model. Remove this trait from the device model and playback commands will work.
The Assistant cannot broadcast messages from one device to the other.
- Make sure to set the device location in the Google Assistant app. Both devices must have the same address in the app.
Still having trouble? See our Support page.