Get AI Transcription Superpowers with Scribe for FreePBX

, ,

I don’t think in the thread asking about the decline in quality and support of the open source project dumping a “hey we have a new commercial module now” link is good form.

2 Likes

What thread are you referring to as the OP title for the thread we are on currently is “Get AI Transcription Superpowers with Scribe for FreePBX” and the category it is currently in is literallyCommercial Modules” ?

HTH: Please consider reviewing the recently added information about muting categories if you do not wish to discuss certain things, such as FreePBX Commercial Modules like Scribe.

Yeah, I mixed up which thread I was in when I replied. My mistake. See, that isn’t so hard. To admit when you’ve made a misstep.

Thanks but I’m well aware on how to mute categories, threads and people.

2 Likes

Has anyone tried this?
@penguinpbx It would be useful if there was a free month especially as it a new module release

Sangoma uses Scribe internally in production.

There will be some Scribe licenses given away in the webinar on December 3rd.

That’s not what @yahdie was asking and when looking at the module I thought the same thing. It looks great but honestly there are a ton of AI solutions out there that over promise and under deliver. It’s just a fact of life with new tech.

It would be really nice to be able to have at least one month to test the functionality. As a Sangoma partner it would be even better to be given a 1 month test lic for all commercial modules.

1 Like

Why not full lic for partners. We are there to promote any Sangoma products. So, the cost for Sangoma should be ridiculous. Globally, we could do some demo and not for using all modules everyday personnaly.

Regarding the other kind of product elsewhere. Brands offer a 2 weeks or 1 month lic for tests for any new client.

1 Like

So what does this get us when using the Scribe module? Since this module is powered by Deepgram, who’s own ToS/AUP must be agreed to in additional to all of Sangoma’s stuff and Deepgram is solution with a cost for use but they do offer a free tier.

So is the Scribe module setting up the users with a free tier account at Deepgram? Some sort of PAYG setup? Who is being paid for this usage/plans? Are we setting up an account with Deepgram or is Sangoma running some sort of self-hosted system with Deepgram on it?

I see Scribe has a cost calculator for usage so what is the deal with that?

Also why are all the screen captures of the Scribe module so small? You can’t open them and see the details of the capture clearly. So, awesome on that. Sales did a great job of making sure the slicks/manuals were clear and readable.

3 Likes

Correct, it’s not really readable. You are right!
Maybe a quick video for demo would be welcome.

There is a webinar scheduled for Dec 3, 2024

1 Like

So they launch the marketing for Scribe on Nov 12th and we can buy it now. Why do we need to wait three weeks to have questions answered?

1 Like

No false!

Hi @BlazeStudios

Thanks for highlighting the image issue, we have fixed it.

also, regarding your other queries, created FAQs section in the wiki to capture such queries/feedbacks. Please refer to the wiki

Please keep sharing the feedback or queries to either improve the module or wiki/documentation.

Thanks
Kapil

That wiki…

Q: Does FreePBX upload call recordings to third-party servers for transcription or analysis purposes?

A: We securely call our partner’s API over HTTPS to perform the transcriptions.

Would be nice if you did something like “Yes, your audio is sent to a third-party server.”

It would be odd to send audio over https… is this web socket ?

1 Like

You can send files over an RESTful API. It is how I do voicemail transcripts.

That would be great for voicemail… and horrible for realtime captions.

1 Like

The module doesn’t do realtime. It transcribes voicemail and call recordings. Nothing real time about it.

You’re absolutely right. How completely non-cutting edge.

1 Like

I guess it depends on how you look at it. This is something that has been requested for some time. It’s a shame there wasn’t an open source option with a commercial add-on but at least it fulfills a need.

Perhaps this means voice recognition integration can be coming down the road.

Real-time transcription was introduced at Astricon. Do we have something like this in production yet?