we have generated over 1 million in revenue for a client in US using 1 voice agent on livekit (case study + client interview coming soon) now why i tell you this is because to show that we dont play around with toy bots and build extremely reliable voice agents on livekit that are killers in production. we built our own voice ai stack on top of livekit, next js, supabase and hosting servers on digital ocean. soon we will open source the entire code and currently looking for agency owners to try out the platform and give us rapid feedback so we can fix bugs before open souring the entire code for public access. the stack does not just include UI that looks like vapi/retell , it has many other parts like call concurrency management server, livekit manager server that gives you api routes to easily start calls, receive end call reports and much more. our goal with this open source project is to completely replace vapi/retell + cekura + chatdash and have one stop shop for voice ai devs to build and deploy agents for clients. here are some internal docs and diagrams that give you overview of the voice ai stack we have built that we use for our clients https://github.com/VocalGrid/diagrams comment below if interested (please do not DM me)

Our team just deployed Qwen-Omni3's S2S model on our infrastructure and it is available today w/o a waitlist. I'd love to get feedback and how you think this will impact the world of chainer-apps and if S2S is actually the future. What do you think? https://models.hathora.dev/model/qwen3-omni

Our team at Hathora, plus Anthropic and Lovable and a few other folks are hosting a Hackathon in NYC at Columbia B-School in partnership with their AI Club. If you're in NYC, we'd love for you to join! Sign up here: https://luma.com/wr3v4hl2

Hey everyone 👋 I’m trying to figure out how to build my own voice AI agent, something similar to VAPI Retell..., but custom for my language and accents. We have over 40 accents, and I want to fine-tune a model that can actually understand and speak like people from each region — not just transcribe words, but really adapt to local pronunciation, tone, and vocabulary. I have experience in full-stack development (React, Spring Boot, Go, etc.), but I’ve never done fine-tuning or building voice pipelines before. So my main questions are: - 🧠 What’s the best way to start building a custom voice AI like this? - 🗣️ How can I collect and fine-tune on many accents effectively? - ⚙️ What models, tools, or frameworks do you recommend for speech-to-text, text-to-speech, and conversation handling? - 💾 Do I need to build my own dataset, or can I use / adapt open ones? If anyone here has worked on custom speech AI or VAPI-like systems, I’d really appreciate some pointers 🙏 Thanks in advance!— Khaled Rouissi 🚀

Hey folks, I just launched a voice AI community on X. While I feel like this is an awesome place to go deep, this feels like a good spot to share things that you're working on in voice AI and maybe even bring some more folks to this community. Check it out! https://x.com/i/communities/1989305815254069555