Adding A.I Feature for Recordings module

Hello everyone.

I am adding a feature for the Recordings module.
This feature allows to add a TTS audio file from elevenlabs.io.
We can enter a text to convert to an audio file by selecting a voice (fr, en and es for now) and the API will download the audio file and convert it for the Recordings module.

You just need to create your account on https://elevenlabs.io and generate your API key and add it on the module.
You can start with a free version, or submit a service for $5/month.

I coded this feature as a pilot, so if anyone else wants to add another one, you can do it easily.

For now, I am testing on 16 and then, I will test on 17.

Coming soon.

6 Likes

Nice! Thank you for that.

Would you consider also adding support for OpenAI Voices? Thanks

Why not. If I can get an API key for free, then yes.
However, if I need to buy a license for each AI just for dev… I think not.
Otherwise, you need to give me a API key for test and dev.

I tested on 17, looks good.

1 Like

So if I understand the enhancement correctly it’s just to be able to create System Recordings directly in FPBX as opposed to creating them outside, then going to system recordings, uploading them etc??

1 Like

Correct.
You could write a text and the audio file will be downloaded to the Recordings module.

2 Likes

Where can I test this?

Well. there is no comit yet.
I’m busy for now.
I will see that the next week maybe.

I raised a PR on 16.
Kapil is off maybe for 2 days…
So, be patient.

Hi.
I ran some tests.
Looks good.
I will wait a while to raise the PR jus to be sure this is no bugs.

I will update you soon.

Many thanks. Let me know if you want me to test.

1 Like

I raised a new PR.
Watting @kgupta to validate the job, and merge it.

See you later.

1 Like

Where was this PR raised at? I don’t see a ticket for this or a PR against the IVR or TTS modules.

Recordings Module

Oh duh. I was conflating this with the AI IVR stuff being talked about.

No, it’s a new feature to create any audio files from text using Recordings module.

It’s a good idea, do you think sangoma would approve it being merged in the main recordings module? I would expect this to be an addon module separate from system recordings.

Yes, of course.

The version is out.
Recording module Version 16.0.17.1

3 Likes

On QA testing obviously.