I used to run a voice ai agency, (from 2023–) and we used vapi, sythflow, elevenlabs, and built our own truly agentic features on top, like memory, email, sms, scaled to 50k/mo but since we only worked with mid-market/growing companies, with 100+ employees, our cost per min became nearly unbearable. It became increasingly difficult to be profitable, stay price competitive with all of the new agencies, and acquire new customers at the same time. ...We decided to invest nearly everything we had into our own Voice AI platform so we could cut down our cost by 5x. To make it cheap, AND fast at the same time, we had to cut out every single middle man. Each white-label or outsourced layer increases latency, and of course COST. Even the biggest platforms like Vapi and Retell are white-labels. Vapi being a white-label of Pipecat, and Retell being a white-label of LiveKit. Not to mention other layers... - ElevenLabs TTS, - Deepgram ASR - + LLM All of which shoot up the price incredibly. ElevenLabs itself costs these platforms a minimum of 0.03/min So what we had to do is use our OWN TTS, a faster ASR and our OWN fine-tuned LLM. (the ASR being NVIDIA RIVA) The result: Voice AI at 0.03/min, that we offer to other agency owners so simply just profit 4x more with what you're already doing with Voice AI! ALSO our API is backwards compatible with nearly all voice AI platforms -so saving 4-5x on our voice ai bill is as easy as 1 line of code in your n8n automation.

Let's get to know each other! Comment below sharing where you are in the world, a career goal you have, and something you like to do for fun. 😊