Sharing a small but exciting milestone from my learning journey 🚀 Over the last few weeks, I’ve been deep into voice infrastructure and SIP, and I finally have a full working setup: 📞 Local Phone → SIP (from Signalwire) → FreeSWITCH → Voice Agent → Response back to the caller The FreeSWITCH server is running on a Debian server on DigitalOcean, and everything is now talking to each other smoothly — SIP, RTP, real-time audio, and AI responses. I’m currently working with a client, and initially we’re aiming to scale this setup to handle ~100 concurrent calls, which is pushing me to really understand: - SIP call flows - Audio streaming - Server performance & scaling - Latency trade-offs vs managed platforms Honestly, this stuff is challenging but insanely exciting. Every time a real phone call hits the server and the agent responds correctly, it feels like magic — but the earned kind 😄 Just wanted to share this win and keep building. If anyone here is working on voice agents, SIP, or FreeSWITCH — would love to connect and exchange notes 🤝

Hi All, So, as I've told you all a few weeks ago, I've published an open source tool called OPBX - which is a business PBX system, that works on top of Cloudonix and provides some interesting capabilities when working with voice agents. I've added multi-tenant capabilities to it - so if you install it, you can use it to service all your customers. At the same time, I'll be launching a SaaS version of OPBX, completely FREE of charge, so that you can use it and build with it. Next week, I'll be holding a special OPBX training session, showing how to integrate OPBX with VAPI, Retell, etc. In addition, I'll show how to build multi-agent IVR trees, warm transfers that work as they should and more. Cloudonix Velocity Training Registration - https://us02web.zoom.us/meeting/register/6D63tRaYSDihkJtUlpNp-A#/registration OPBX Github Repository - https://github.com/greenfieldtech-nirs/opbx Looking forward to seeing you all. Nir S

I'm currently handling about 6k calls per day between a few different enterprises. We're likely going to implement and build this all out in LiveKit but I also need a better eval + simulation solution and I don't really want to build it out as well. I have seen Coval but I don't know the cost yet -- what else it out there that you guys use that can give me actionable feedback of missed tool calls, hallucinations, etc.

While wrappers are great for MVPs, building your own orchestration layer gives you 𝗳𝘂𝗹𝗹 𝗼𝘄𝗻𝗲𝗿𝘀𝗵𝗶𝗽, 𝘀𝗶𝗴𝗻𝗶𝗳𝗶𝗰𝗮𝗻𝘁𝗹𝘆 𝗹𝗼𝘄𝗲𝗿 𝗰𝗼𝘀𝘁𝘀, 𝗮𝗻𝗱 𝗴𝗿𝗮𝗻𝘂𝗹𝗮𝗿 𝗰𝗼𝗻𝘁𝗿𝗼𝗹 over the entire conversational pipeline. I designed this engine to fully replace third-party wrappers like Vapi & Retell AI. Here is a deep dive into what’s under the hood: 🔄 𝗗𝘆𝗻𝗮𝗺𝗶𝗰 𝗔𝗴𝗲𝗻𝘁 𝗖𝗼𝗻𝗳𝗶𝗴𝘂𝗿𝗮𝘁𝗶𝗼𝗻 (𝗥𝗲𝗮𝗹-𝗧𝗶𝗺𝗲 𝗛𝘆𝗱𝗿𝗮𝘁𝗶𝗼𝗻) Hardcoding agents is a trap. I implemented a system that executes an API call upon call initialization. • 𝗛𝗼𝘁-𝗦𝘄𝗮𝗽𝗽𝗮𝗯𝗹𝗲 𝗣𝗲𝗿𝘀𝗼𝗻𝗮𝘀: A single engine instance can instantly apply unique System Prompts, Voice IDs, and Temperature settings based on backend parameters. • 𝗥𝗲𝘀𝘂𝗹𝘁: You can power thousands of unique agents (e.g., specific to different businesses) without ever redeploying the core code or creating a new instance. 🛠️ 𝗖𝗼𝗻𝘁𝗲𝘅𝘁-𝗔𝘄𝗮𝗿𝗲 𝗙𝘂𝗻𝗰𝘁𝗶𝗼𝗻 𝗥𝗼𝘂𝘁𝗲𝗿 When building raw infrastructure, manually mapping tools to agents is a major architectural hassle. I built specialized helper logic for 𝗗𝘆𝗻𝗮𝗺𝗶𝗰 𝗧𝗼𝗼𝗹 𝗜𝗻𝗷𝗲𝗰𝘁𝗶𝗼𝗻 to solve this. • 𝗠𝗼𝗱𝘂𝗹𝗮𝗿 𝗟𝗼𝗴𝗶𝗰: The router decouples the orchestration engine from business logic. It parses the backend setup and assigns only the specific tools defined in that agent's configuration (e.g., loading "Appointment Booking" tools only when the specific use-case demands it). 💾 𝗗𝗮𝘁𝗮 𝗣𝗲𝗿𝘀𝗶𝘀𝘁𝗲𝗻𝗰𝗲 & 𝗣𝗼𝘀𝘁-𝗖𝗮𝗹𝗹 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 Logs aren't enough. I built a save_conversation function that aggregates the full session payload and triggers intelligent sub-functions immediately after the call: • 𝗖𝗮𝗹𝗹 𝗦𝘂𝗺𝗺𝗮𝗿𝘆: Generates a natural language recap via LLM. • 𝗖𝗮𝗹𝗹 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗶𝗼𝗻: Structurally classifies the outcome (e.g., "Booked", "Inquiry", "Failed"). • 𝗧𝗲𝗹𝗲𝗺𝗲𝘁𝗿𝘆: Captures precise Token Usage (for billing) and Latency statistics alongside the transcript. 🛡️ 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗚𝘂𝗮𝗿𝗱𝗿𝗮𝗶𝗹𝘀 To prevent runaway costs and "zombie" connections, I engineered active background monitors: • 𝗜𝗻𝗮𝗰𝘁𝗶𝘃𝗶𝘁𝘆 𝗠𝗼𝗻𝗶𝘁𝗼𝗿: Detects silence (30s default) and gracefully terminates the session.

Hello all, I'm trying to investigate a few hypothesis I have regarding the AI Voice Agent market. My questions are mostly related to security, observability, billing and load management. In order to do so, I've built the following Google Form: https://forms.gle/oFeM9J9WV9DRX9267 If you could please answer it, I would highly appreciate it - also, once I have all the data compiled - I will publish a post with all my findings, so that people can learn from this study as well. Much Appreciated.