As the AI hardware, software and pre-trained models become more and more abundant, advanced and accessible, we see the growing need to have AI-centric edge real-time inference tool, allowing one to build very own custom real-time inference system to be easily integrated into existing RTC (SIP, WebRTC, MQTT) infrastructure. Infernos strives to fill in this role by providing a platform for running all kinds of models (TTS, STT, image classification etc) with a very little integration effort.