The WebRTC.ventures Blog.

Telephony agents built on Voice AI platforms like Retell AI make it easy for businesses to run live phone conversations through an AI agent instead of an IVR or human call center staff. Getting one of these agents into production means solving two things: integrating it with existing telephony infrastructure, and adding a compliance layer that preserves the organization's security and governance standards. On the integration side, the agent must connect to existing PBXs, SIP

How can AI scale clinical roleplay training beyond what live sessions with human trainers can deliver? For CETA Global, a nonprofit that trains frontline mental health providers around the world, the answer was an AI “flight simulator”: a real-time platform where psychologists practice difficult client sessions with an AI-generated patient, complete with a lip-synced streaming avatar, live coaching, and replayable feedback. Behind the scenes, a team of six AI agents works in concert to play the

An agentic workflow is the disciplined form of what's often called "vibe coding": giving a coding agent a task and the context it needs, then letting it plan, edit, and test across your codebase while you review the results. When it's done well, the productivity gains can be dramatic. In his vibe-coding WebRTC post, Philipp Hancke describes getting changes merged into libWebRTC, Firefox, and even Pion, work that used to wait on finding a free

Most AI applications operate in turns: a user submits input, the model processes it, and a response is returned. A delay of a second or two goes unnoticed. Live voice AI does not offer that margin. In a roleplay training session where a psychologist practices a difficult client conversation with an AI-simulated patient, every unnatural pause undermines the realism of the exercise. And with it, the training value. This was the central engineering challenge behind

The Janus WebRTC Server is one of the most pleasant WebRTC media servers to build on, and it’s a bonus that it is open source. It's small, fast, and modular: a thin C core with plugins for exactly the workloads you need, a clean signaling API, and a videoroom plugin that handles SFU-style conferences efficiently. However, moving beyond one instance and spreading rooms across a pool introduces its own challenges. Unlike some media servers, Janus

WebRTC.ventures recently completed a Kurento to LiveKit migration for a client running one of the more complex real-time platforms we've worked on, and we did it without downtime. The platform had conferences, webinars, inbound telephony (call center), outbound telephony, recordings, and live transcription and translation in English and Spanish — all in production, serving real users every day throughout the entire migration. This post covers the key decisions and findings. The full technical case study,

Vonage recently released the Video Connector SDK and a companion Pipecat Transport. These new tools let a server-side AI agent join a Vonage Video session as a full participant, with real-time audio and video. Together they make it possible to build an AI Avatar that doesn't just chat, but actually listens and speaks inside a live video call. That shift matters because most AI interaction today is still text-based — a chatbot as the first

Building real-time applications with AI requires disciplined engineering, not just clever prompting. That was the core thesis of the June 2026 AI Tinkerers San Salvador meetup, sponsored by WebRTC.ventures and AgilityFeat. In this post, we share key takeaways from the event: starting with real-world applications like an offline-first AI application for sign language translation, then we dove into the mechanics of LLMs, the power of Spec-Driven Development, and a case study on using AI to

“Real-time voice AI only works when infrastructure makes latency feel invisible. For us, that meant changing the shape of our WebRTC deployment without changing what clients expect from WebRTC itself.” A firsthand look at how OpenAI rearchitected their entire WebRTC stack for global scale, straight from Yi Zhang, co-author of their deep-dive post, How OpenAI delivers low-latency voice AI at scale. He was part of the team behind this overhaul, addressing pauses, interruptions, and delayed barge-in for

If your call center wants to use AI to automate call intake and handoff without losing the personal touch, a SignalWire Voice Agent can help. In this tutorial, you’ll build a Python-based AI receptionist that greets callers, captures the information you need, qualifies intent, and hands the call off to a human at the right moment.  The SignalWire Agents SDK makes orchestration across telephony, speech processing, and call control much easier to manage. That means

Voice agents are moving out of demos and into production systems that handle customer requests, account actions, healthcare questions, financial details, and support escalations. That changes the security model. A voice agent is listening, deciding, acting, and speaking in real time. That is why voice agent security cannot rely only on prompts or post-call transcript review. Sensitive data guardrails need to run inside the live pipeline: before sensitive data reaches the model, before tools are

Voice AI systems generate more than recordings and transcripts. Every production interaction produces a web of artifacts across multiple systems: call-setup metadata, ASR output, LLM responses, tool calls, CRM updates, escalation events, and compliance-relevant signals like caller identity verification. Most Voice AI architectures store some of these. Few store all of them in a way that survives an audit, a customer dispute, or a platform migration. That gap has a name and an emerging solution.

Choosing to go open source over a CPaaS for your WebRTC media stack is a strategic decision about control, flexibility, and long-term ownership. For teams building real-time products, an open source WebRTC media server can offer the freedom to customize media handling, integrate deeply with your architecture, and avoid the constraints of a managed platform’s pricing or roadmap. “With great power comes with great responsibility,” which is why open source is rarely the easiest path.

We built a Live Sales AI Presenter to model what a real-time AI avatar system looks like when it operates inside a real workflow rather than a demo sandbox. It is a slide-aware AI sales presenter combining deck ingestion, presentation control, live Q&A, WebRTC media, Pipecat orchestration, OpenAI Realtime, and HeyGen live avatar. You can watch the walkthrough here and the code is here. What the prototype makes visible is that a real-time AI avatar

A year ago, the question was whether AI avatar realism was convincing enough to use in front of a customer, whether for sales, support, or onboarding. That question is largely settled. These digital presenters have moved past the novelty stage and are now regular players in the Voice AI stack. HeyGen, Synthesia, D-ID, and Colossyan produce polished talking-head video with low enough latency that avatar output is no longer the primary differentiator. Today’s challenge is

Porting your business phone number to add AI call handling is an operational risk most companies don't need to take. Your number may be tied to other services. Your carrier contract may have obligations. And if something goes wrong mid-port, customer calls go nowhere. SIP forwarding sidesteps all of it. Your carrier forwards calls from your existing PSTN number to a SIP endpoint. The AI agent picks up, handles the conversation, and transfers to your

WhatsApp has over two billion active users, but for developers building real-time communications applications, it’s long been a walled garden. That’s changing. WhatsApp Business Calling is opening new doors for SIP-based integrations, and the technical landscape around it is evolving fast. In this episode, Dan Jenkins joins us to break down what combining WhatsApp Business Calling with WebRTC and SIP actually looks like in practice: the use cases, the architecture decisions, and the technical challenges that come with it.

Peermetrics is WebRTC.ventures' open-source WebRTC monitoring stack. Earlier this year, a client pushed it to a scale that stress-tested assumptions you can't easily replicate in development: thousands of video conferences a day totaling over a million events. That traffic surfaced things that only show up at volume. The result: Peermetrics SDK v2.8, @peermetrics/webrtc-stats v5.9.0, and a round of API and dashboard changes focused on server-side aggregation, Redis caching, cache pre-warming, and drilldown correctness. We also

Prompt engineering gets you a demo. Context engineering gets you a production Voice AI agent. Think of LLMs as the world's most brilliant librarians: they've read almost everything ever written, but without your help, they have the short-term memory of a goldfish. For text-based chatbots, a forgetful LLM is annoying. For a real-time voice agent, it kills the experience entirely. This is because the goal for Voice AI is human-level fluidity with sub-second latency. To

AWS autoscaling works well for stateless applications. And for the stateless components of VoIP and real-time systems like APIs and routing backends, AWS Auto Scaling groups (ASGs) and Amazon Elastic Container Service (ECS) do exactly what they're designed to do. But stateful VoIP infrastructure components like FreeSWITCH and RTPEngine maintain active SIP sessions and media streams tightly bound to specific instances. Because each instance holds live call state, scaling events must be carefully coordinated. Otherwise,