Gladys Assistant 4.33: Merry Christmas to all šŸŽ„

News
1

Hello everyone!

This is the last release of the year, I present to you Gladys Assistant 4.33 :santa_claus:

On the agenda, a bunch of fixes and above all a feature everyone was waiting for: the ability to make Gladys speak :speaking_head::speaker_low_volume:

This is a first alpha, and Iā€™m really happy with the result: Gladys speaks in any language, with a smooth (and feminine!) voice, which took me a long time to choose :slight_smile:

This feature is available via Gladys Plus only, because the TTS API I use is paid.

Iā€™m offering it for now at no extra cost during the alpha (Iā€™ll check the bill at the end of the month and decide), I set a limit of 200 voices generated per month per account to avoid unpleasant surprises on the bill.

If you ever reach the limit, tell me ā€” I can reset it without problems or change it if it makes economic sense.

This feature only runs on Sonos for now since we only manage Sonos speakers, but Iā€™m not opposed to adding it to other speaker brands if they have an API that allows it :slight_smile:

Itā€™s an alpha so be indulgent if itā€™s not perfect :wink:

Note: If you have already added your Sonos speaker to Gladys, you need to remove it and add it again (in Gladys) to create the new feature on the device!

Improvements/fixes

  • Sensor states purge now happens at 4:00 AM instead of at minute 4 of every hour due to an error in the CRON!
  • Zigbee2mqtt: The Ā« child lock Ā» is now functional :slight_smile:
  • Zigbee2mqtt: Handling of raw Ā« illuminance Ā» (not in lux) for devices that report it
  • Documentation: Added a link to all documentations for each integration, and fixed a crash on the Philips Hue page
  • Added precise error messages in the interface and attempted fix for a tablet disconnecting in locked alarm mode (your bug @spenceur)

Thanks to @Brisou, @AlexTrovato and @_Will_71 for their contribution in this release :raising_hands:

The full CHANGELOG is available here.

A quick word before Christmas

Today was my last Gladys day before Christmas :slight_smile: Iā€™m at a freelance client for the rest of the week, and then at the end of the week Iā€™ll shut my laptop down until January 2.

Iā€™ll be back to Gladys on Friday, January 5.

In case of absolute emergency, Iā€™m available this week if 4.33 introduces critical bugs. Iā€™ll take the time the next evenings to check for reports.

For improvements / new features, that will be in January :slight_smile:

Happy holidays to everyone :christmas_tree:

Replay of Sundayā€™s live coding

For those who missed it, the replay of my live coding Ā« A Z-Wave integration in one day Ā» is available on YouTube ā†’ https://www.youtube.com/watch?v=f6mWvy2kWSs&t=461s

Thanks to those who accompanied me during this 7.5-hour live!!

How to update?

If you installed Gladys with the official Raspberry Pi OS image, your instances will update automatically within the next hours. This can take up to 24h, donā€™t panic.

If you installed Gladys with Docker, make sure you are using Watchtower (See the documentation)

13 Likes
2

Happy holidays to you too :partying_face:

1 Like
3

Thanks for all that, canā€™t wait to tell my kids to sit properly at the table via Sonos hahaha !

Merry Christmas :christmas_tree:

4 Likes
4

I highlighted the TTS feature on Twitter and Reddit!

Twitter: https://twitter.com/pierregillesl/status/1736790862615421305

Reddit on /smarthome: https://www.reddit.com/r/smarthome/comments/18ldjst/finally_my_own_jarvis_at_home/

Donā€™t hesitate to support!!

1 Like
5

[quote="Philou,

2 Likes
6

@pierre-gilles
I donā€™t have a Sonos yet, can I resell my 200 messages? :roll_eyes: :man_running:
Well, thereā€™s still time to order :santa: a Sonos One for Christmas! :blush:

7

Yes I really recommend these speakers, insane quality and powerful sound. The quality/price ratio is really great, especially if

5 Likes
8

Thatā€™s funny, I listened with the monitorā€™s speakers and I thought I heard in the distance Ā« whoā€™s that fā€¦ English woman youā€™re talking to??? Ā» Itā€™s true the sound is good!!! :santa: :rofl: :rofl: :rofl:

1 Like
9

Great job, I donā€™t have a speaker but it really looks great in the video!

Out of curiosity, which API are you using for the TTS (text-to-speech)? OpenAIā€™s, like for the chat?

10

No, OpenAIā€™s worked really well in English but not very well in French.

I went with ElevenLabs, theyā€™re the leaders in AI-generated voice :slight_smile: For the most part, the Ā« mastodons Ā» (AWS, Google Cloud) seem a bit behind on the subject, I think ā€” the rapid advances in AI over the past few months havenā€™t yet been reflected in their TTS

1 Like
11

Related information if you have a little time:

And happy holidays to everyone.

2 Likes
12

@pierre-gilles
Concretely, a text is sent and it returns an audio file that just needs to be broadcast, is that right?

13

The most basic option would be to use the local jack port, to allow using a non-connected speaker :slight_smile:

Happy holidays to everyone, enjoy!

14

I see the TTS credits I prepaid are melting away like snow in the sun :joy:

Thatā€™s cool, it means the feature is popular!!

If it ever cuts out at some point, it might be because Iā€™ll need to move to a higher plan, so donā€™t panic ^^

Why not, but Iā€™m a bit skeptical about that use ā€” who would actually do that in their connected home?

That already means the Gladys box would have to be where the speaker is located, that multi-room isnā€™t possible, and that the speaker connected to Gladys would have to be truly dedicated to voice only (no music).

I get the feeling itā€™s a bit of a hack thatā€™s not great and wonā€™t really be used :sweat_smile: From experience, for example on a Pi, the jack port is really not good and the sound will be very weak unless itā€™s a speaker with a whole amplifier.

Edit: Shared on LinkedIn

2 Likes
15

@pierre-gilles

Butā€¦Butā€¦Butā€¦Who took my share ??? :japanese_ogre:

Itā€™s true that this shows itā€™s a requested feature, but ideally it would be to have an offline solution ā€” would it be possible later to see if Gladys could be interfaced and run with containers designed for that, like for Node-RED?

TTS

Speech recognition
https://korben.info/whisper-interface.html

Otherwise, in case I forget or a meteorite falls on me Ā« Merry Christmas and happy end-of-year holidays to everyone, hoping that 2024 will be richer in announcements on the Gladys side and in number of users! Ā»

16

We could connect a Bluetooth transmitter to it, to hook up a speaker of the same name (even more DIY, but itā€™s true that I thought about it) while waiting for Sonos.
Happy end-of-year holidays to everyone :wink:

17

[quote="cce

18

And to think I didnā€™t take the time to implement it ^^ā€™

Iā€™ll certainly drop by when I have time for this new action

19

I think itā€™s the initial excitement. Ā« Yay, a new toy! Ā» but as soon as my wife tells me, Ā« Okay, isnā€™t that enough of your nonsense? Ā» from having to hear Ā« oh dear, that catā€¦ Ā» every time

1 Like
20

There are less resource-hungry solutionsā€¦ :roll_eyes: :innocent:

https://hub.docker.com/r/alphacep/kaldi-vosk-server

it would be really nice to have voice on Gladys in 2024! :christmas_tree: