Squawk - Real-Time Local Text-to-Speech with AI

Squawk - Real-Time Local Text-to-Speech with AI v0.0.4

royshilkrot

Member
royshilkrot submitted a new resource:

Squawk - Real-Time Local Text-to-Speech with AI - Generative AI engine for speech in

The OBS Squawk plugin adds powerful voice cloning capabilities to OBS by leveraging sherpa-onnx. With this plugin, you can generate speech on the fly and in real-time inside OBS without any external services or access to the network.

If you like this work drop a review here or ⭐ the Github repo, support our work on Github, Patreon or...

Read more about this resource...
 

Tyr808

Member
Combining this with the OCR capabilities of your other plugin, I can see this possibly being a very realistic workflow to adding voice acting to an old text only RPG for example. There's definitely fine tuning involved there, but it looks like the core functionality is there. neat stuff!

Definitely keeping an eye on this project
 

MartinBrieger

New Member
I end up with a path like this on UNBUNTU
/home/runner/work/obs-squawk/obs-squawk/build_x86_64/release/Release/lib/x86_64-linux-gnu/obs-plugins/

Not very intuitive .....
 

ASchneider

Member
Hello @royshilkrot, thanks for creating this very useful plugin. Do you think there's a way of using it with Streamer.bot? Like sending a TTS request directly from it, maybe using websocket, OBS RAW or C#? Thanks!
 

royshilkrot

Member
Hello @royshilkrot, thanks for creating this very useful plugin. Do you think there's a way of using it with Streamer.bot? Like sending a TTS request directly from it, maybe using websocket, OBS RAW or C#? Thanks!
easy - just write to a text file somewhere and the plugin source will monitor it for the change. it will generate speech as soon as there's a change
 

only_ripper

New Member
is it just me when using audio monitoring i get weird audio cracking noise, on stream sounds fine but can't make it sound the same when hearing it, tried on 2 different obs and that's the only source where i keep facing this problem.. Anyway, this is a great plugin overall ✌️
 

royshilkrot

Member
is it just me when using audio monitoring i get weird audio cracking noise, on stream sounds fine but can't make it sound the same when hearing it, tried on 2 different obs and that's the only source where i keep facing this problem.. Anyway, this is a great plugin overall ✌️
Yes there's a problem in the audio when monitoring
I'm still trying to find the bug..
 

royshilkrot

Member
This looks amazing. Is there by chance websocket support to switch voices?
The voice choice is part of the settings of the source which are editable from the outside with e.g. Advanced Scene Switcher
I'm fairly certain streamer.bot can do it as well

If you're unsuccessful in any way - let me know and I'll try to figure it out
 

Blootrix

New Member
Hey! First off, I absolutely love this plugin, it's fantastic!

I have two questions about the plugin if you don't mind.
Is there a way I can use a voice that I have trained?
Is there a way I can have it read messages from a URL source like Ko-Fi or StreamElements?

On the subject of the voice crackling when monitoring audio, is the audio being generate 16bit 48khz? Just wondering if maybe that's the issue.
 

only_ripper

New Member
Hey! First off, I absolutely love this plugin, it's fantastic!

I have two questions about the plugin if you don't mind.
Is there a way I can use a voice that I have trained?
Is there a way I can have it read messages from a URL source like Ko-Fi or StreamElements?

On the subject of the voice crackling when monitoring audio, is the audio being generate 16bit 48khz? Just wondering if maybe that's the issue.
Hi,
Best way you can use it is with streamer.bot, there is basically no limits in this way, have the text generated to a .txt and set the plugin to read from that, you can use it for those two and for many more stuff, like First Words from a viewer per stream to greed them with a message for example, as long as you can output text to a file, you can do everything you want (I also have that .txt clear itself using c# after a couple of seconds so it doesn't stack up and repeat the old text), the only downside at the moment is the cracked sound, unfortunately.
 

Blootrix

New Member
Hi,
Best way you can use it is with streamer.bot, there is basically no limits in this way, have the text generated to a .txt and set the plugin to read from that, you can use it for those two and for many more stuff, like First Words from a viewer per stream to greed them with a message for example, as long as you can output text to a file, you can do everything you want (I also have that .txt clear itself using c# after a couple of seconds so it doesn't stack up and repeat the old text), the only downside at the moment is the cracked sound, unfortunately.
Thank you very much for the reply. I'll check out that bot you mentioned just now!

Only thing I need to figure out then is how to use a voice I've generated myself! (And how to generate my own voice too )
 

DJAtomika

New Member
Hi,
I'm trying to get this working with OBS 30.2.3. I'm using the latest version of the plugin through manual installation.

Whenever I try to create a TTS source in OBS, it instantly crashes my program to desktop. Attached is my latest crash report from OBS. I have no idea what's causing the crash nor do I have a way to figure out what. I've tried both the manual install and the .exe installer for Windows 11 64bit and neither work.

Would appreciate some help in figuring out what's going on. Thanks.
 

Attachments

  • Crash 2024-08-20 13-49-22.txt
    311.1 KB · Views: 6

DrSupermanShadow

New Member
Why is it that every time I try to search for an obs tts reader and great Autism-friendly application(uI) that will allow me to read and program my own voice( using a voice reader programer interface and a bit of ai) for free on or off stream,
I get programs that are not even closely related to what I am looking for?

In 2024 TTS reading for streams and obs users should be a standard dock/ feature.

this is not A modern TTS System. It's a manual input system that requires you to create the file to be read on stream...
this says its a 2024 plugin but its already outdated...in terms of TTS.


Linked Google-based TTS or any free precoded reader chatbot into an obs plugin.
that
1. allowsthe user to exclude reading certain chatbots
2. Read chat for users who use the command like !say
3. can be set up to read certain types of users or specific regular followers or subscribers...
4. can be set to activate by bits or channel points or direct donations.
5. can be approved or disapproved by mods.
6. Using the program command center and read program;
you can create your own TTS voices or use the provided ones.

in 2024 this is one of the most wanted obs-type plugins and or features
All of the bots that are currently out there are an external program and usually are not free, and add more strain to otherwise already tight budgets( they also stink in terms of UI friendliness) >
I personally, have been searching for more than 3 years for this great disability accommodation feature for OBS.

Btw, this topic is the first search result on Google in 2024. when you type TTS chat reader plug in for obs.


The info below: is for those who feel the need to be what is defined as a forum troll after I post.

and to those about to comment about cost:
There are more people than you know about
who want this country's way of doing programs and software (and basically running the world economy). it's in the billions FYI.
An autism-based person has written out this post, your definitions of if this post is acceptable must be with this Nation's Autism Law Category.
PS. I can post this country's flag and or public. P.E.C information panels into any forum post
no TOS or any contract can override it. It's a PUBLIC INFO LAW.
It also allows the poster to only be subject to NHN's Forum laws and Code of Conduct laws( or any NHN laws really) as of 2022 NEW HORIZENS NATION FLAG Cover.png
 

only_ripper

New Member
Why is it that every time I try to search for an obs tts reader and great Autism-friendly application(uI) that will allow me to read and program my own voice( using a voice reader programer interface and a bit of ai) for free on or off stream,
I get programs that are not even closely related to what I am looking for?

In 2024 TTS reading for streams and obs users should be a standard dock/ feature.

this is not A modern TTS System. It's a manual input system that requires you to create the file to be read on stream...
this says its a 2024 plugin but its already outdated...in terms of TTS.


Linked Google-based TTS or any free precoded reader chatbot into an obs plugin.
that
1. allowsthe user to exclude reading certain chatbots
2. Read chat for users who use the command like !say
3. can be set up to read certain types of users or specific regular followers or subscribers...
4. can be set to activate by bits or channel points or direct donations.
5. can be approved or disapproved by mods.
6. Using the program command center and read program;
you can create your own TTS voices or use the provided ones.

in 2024 this is one of the most wanted obs-type plugins and or features
All of the bots that are currently out there are an external program and usually are not free, and add more strain to otherwise already tight budgets( they also stink in terms of UI friendliness) >
I personally, have been searching for more than 3 years for this great disability accommodation feature for OBS.

Btw, this topic is the first search result on Google in 2024. when you type TTS chat reader plug in for obs.


The info below: is for those who feel the need to be what is defined as a forum troll after I post.

and to those about to comment about cost:
There are more people than you know about
who want this country's way of doing programs and software (and basically running the world economy). it's in the billions FYI.
An autism-based person has written out this post, your definitions of if this post is acceptable must be with this Nation's Autism Law Category.
PS. I can post this country's flag and or public. P.E.C information panels into any forum post
no TOS or any contract can override it. It's a PUBLIC INFO LAW.
It also allows the poster to only be subject to NHN's Forum laws and Code of Conduct laws( or any NHN laws really) as of 2022View attachment 107511
I see that you are pretty mad everywhere about a lot of stuff, like people which does plugins must keep their plugins up to date and work on each aspect to make them usable for everyone.
Most of the plugins are free (almost all of them), while being open source as well, each plugin takes time to make and even more to keep it updated, almost every plugin was made by someone with the intentions to make creators live easier and more enjoyable, while earning absolutely nothing, demanding them to work and make the perfect plugins is nonsense, we should appreciate everything was already made by the OBS Community towards us, again, for free.
Personally, I think everyone should be grateful those plugins exist in the first place, and we should thank them, your anger is not with them.
 
Top