Harry789

Dograh v1.20 - Open-source voice AI agent platform (with speech-to-speech support)

https://lemmy.world/post/45470336

Apr 11 at 2:20pmWeb
Show threada6kmeApr 11
Hello There. I am one of the maintainers of Dograh. Please let me know if you have any questions regarding Voice AI Agents/ Self hosting Language Models/ Speech to Speech live models in general. Looking forward to hearing from you all. Thanks!
Show threadhendrikApr 11
Is there any open-weights speech-to-speech model available?
Show threada6kmeApr 11
Hello @hendrik - I haven’t personally tried an open weight S2S model yet. Its in my agenda for the week. I am building a framework like approach to run any open source model at github.com/dograh-hq/speaches and I can add one of S2S models there. Thanks!
GitHub - dograh-hq/speaches

Contribute to dograh-hq/speaches development by creating an account on GitHub.

GitHub
Show threadhendrikApr 11
That sounds great. And I’m always interested in multiple languages as well. I mean Gemini Live can do German and all kinds of languages. But that kind of functionality is lacking in most of the shiny new tech-demos. We have some STT and TTS. But for example Kokoro only does English.