I've been asked to develop an action for either Google Mini
or Amazon Alexa
that allows the device to record the audio from a meeting (no time limit) and send that script to the meeting-participants.
Is it possible to record audio for extended periods of time with these devices?
I'm aware that both the Mini
and Alexa
are 'always recording' but it seems they just overwrite local storage. As an alternative to recording for an extended period, is it possible then to output these recordings to a web-service and concatenate them into one file on the back-end?
Thanks
Can You Set Alexa to Record Conversations? Alexa's default settings record all interactions you have with the device at any time. So it's critical to know Alexa can record conversations, but only after using the wake word. Thankfully, you can go into your Alexa app and access settings to change your preferences.
The companies will tell you that they do not record unless you've activated them with the “Wake Word.” The default for Alexa is “Alexa” and the defaults for Google Home are “OK Google” and “Hey Google.”
The short answer is yes, Alexa is always listening to you. Or rather the microphone on your Alexa smart speaker is always active (by default, at least) and is constantly monitoring voices in your home in order to hear its wake word.
But each platform has its strengths. For a better virtual assistant and smart displays, we think Google's the one to beat, but Alexa dominates in terms of music quality and smart home prowess. Despite these strengths, the one you choose ultimately boils down to personal preference.
First of all, the devices aren't "always recording". They are "always listening". Anything that is stored in memory is for a very short period of time - just long enough to see if it is the triggering hotword and/or to forward anything else said once triggered.
There is no authorized way for you to tap into this mic.
Similarly, neither device gives you access to the audio recorded through a server - they just send you text transcripts. Google does let you listen to what was recorded on their servers, but you can't access it as a third-party.
On the Alexa, there probably isn't much you can do. It won't forward text that doesn't match one of the Intents, and open-ended text is deprecated and discouraged.
With the Assistant, using the Action SDK, you can get most of the text spoken. But there are a few caveats. Although you'll get text, you won't get who spoke the text, nor the timing of things (was it spoken slowly? quickly? was there a long pause between a question and response?). The Assistant will also send things in batches, either when it thinks you've said "enough" or when it thinks you've finished what you're saying, and your server then needs to send back a response (which could be as simple as a very short audio file), but while it is "processing", you'll miss everything that was said.
In short - while possible, the current hardware and system software is unsuited to this sort of application.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With