Thanks, it depends a lot on what are you trying to do, if you are going to use just cloud providers you need need a heavy system , a simple system with 2 cores 4 GB RAM with asterisk should suffice, if the call volume is low.
But if you want to go full local route then you need GPU and if call volume is heavy then more RAM and CPU.
Basic hardware should be handle 20-30 concurrent calls with a cloud provider.
Hi HN folks ! I am the author of AVA, a self hosted AI Voice Agent that plugs into Asterisk/Freepbx so you own all the aspects of an AI Voice agent in your own infrastructure.
It uses Asterisk native Audiosocket/RTP with python engine to run STT,LLM and TTS loop. The project support several full providers openai, gemini, grok, elevenlabs out of the box and also provides options to build custom pipelines by choosing different stt tts and llm.
It also supports full local agent if you have a GPU with 25GB RAM which enables realtime conversation along with tool calling.
I started this as a hobby project last year when I started exploring voice agents and every saw every Saas tried to lock you in their eco system. Since then project has taken off and a lot of asterisk people started using it due to ease of setup and wide range of providers and flexibility.
I am also exploring to convert this into a more polished version for Operators to handle multiple customer and multi tenant environment and I would genuinely like to hear feedback on it.
Happy to answer any questions and gather any feedback. Thanks.
Freepbx, Fusionpbx uses Freeswitch. This is built for Asterisk but there is nothing stopping from setting a SIP trunk from Freeswtich to Asterisk box to make use of Agents like this user did https://github.com/hkjarral/AVA-AI-Voice-Agent-for-Asterisk/...
Freepbxx built on Asterisk primarily for single orgs whereas Fusionpbx is built on freeswitch designed for multi tenant and carrier grade deployments. Is that what you were looking for or something more specific in relation to this project ?
There was a window between the early 2000s and the late 2010s when self-hosted telephony made sense. But these days, there are so many cheap commercial services that it is almost impossible to justify the cost of maintaining your own voice infrastructure. Moreover, the proliferation of spam / scam calls has tarnished the perception phone interactions so much that most people avoid them at all costs.
And the few people who do still use the phone absolutely, unequivocally do not want to talk to an AI voice agent ever.
14 comments
[ 3.0 ms ] story [ 39.5 ms ] threadBut if you want to go full local route then you need GPU and if call volume is heavy then more RAM and CPU.
Basic hardware should be handle 20-30 concurrent calls with a cloud provider.
Full breakdown of hardware requirements is captured in this document: https://github.com/hkjarral/AVA-AI-Voice-Agent-for-Asterisk/...
It also supports full local agent if you have a GPU with 25GB RAM which enables realtime conversation along with tool calling.
I started this as a hobby project last year when I started exploring voice agents and every saw every Saas tried to lock you in their eco system. Since then project has taken off and a lot of asterisk people started using it due to ease of setup and wide range of providers and flexibility.
Core is MIT: https://github.com/hkjarral/AVA-AI-Voice-Agent-for-Asterisk Free and open source forever.
No install Mock UI us hosted at: https://demo.agent6789.com
I am also exploring to convert this into a more polished version for Operators to handle multiple customer and multi tenant environment and I would genuinely like to hear feedback on it.
Happy to answer any questions and gather any feedback. Thanks.
And the few people who do still use the phone absolutely, unequivocally do not want to talk to an AI voice agent ever.