The WebRTC.ventures Blog.

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,

For the past several years, AV1 has been generating real excitement in the WebRTC and real-time communications space. Better compression, no licensing fees, and backing from the biggest names in tech made it look like the codec the industry had been waiting for. Then, on March 23, 2026, Dolby Laboratories filed patent infringement suits against Snap Inc., targeting Snap's use of HEVC and AV1 to encode and transcode videos in its Snapchat application. For developers

For years, developers building real-time video and audio applications on the web have worked under a quiet constraint: whatever protocol you chose, it had to work in Safari.  Well, that’s not 100% true because many WebRTC apps used to have disclaimers like “works best in Chrome”. When Safari finally fully supported WebRTC, it was a happy day for WebRTC developers. We could stop asking users “what browser are you using?” when responding to bug reports.

Your users report poor call quality, a dropped call, or a connection that never got established. But what actually happened? In this episode of WebRTC Live, we’ll break down what commonly fails in production WebRTC apps, how experienced teams debug live incidents, and how to build the visibility that keeps you ahead of problems. Guest host Alberto Gonzalez, CTO of WebRTC.ventures, sits down with Justin Williams, Senior WebRTC Engineer at WebRTC.ventures. Justin brings hands-on experience building real-time

Have you ever noticed a very subtle and easy-to-miss security setting in WhatsApp called “Protect IP address in calls”? Some people scroll past it without a second thought. Others may enable it without fully understanding why it exists. The real question is: Why would WhatsApp even give users the option? If calls are already “end-to-end” encrypted, what exactly still needs protection? VoIP (Voice over Internet Protocol) and WebRTC (Web Real-time Communication) calls are often considered

AI-driven QA testing is reshaping how teams validate real-time applications. Doing it well requires intentional processes, shared knowledge, and a collaborative culture that allows teams to use AI responsibly and consistently. Our WebRTC.ventures QA team has approached this with a clear mindset: the real value of AI in software testing comes from building operational systems that let the entire team benefit, while maintaining strong testing discipline. We've built internal practices that make AI a core

The “no-code” revolution has made AI voicebots significantly more accessible. Non-technical teams can now launch voicebots quickly with platforms like Vapi or Bland AI and start automating customer interactions without a dedicated engineering team. In practice, running a good voicebot requires serious operational strategy. Long-term success depends on factors that aren’t always visible upfront:  how the system handles context how deeply it integrates with your existing tools how it performs under real-time constraints how costs

WebRTC.ventures CTO Alberto González recently joined the Software Defined Talk podcast to share insights on building voice, video, and streaming applications for enterprise use. In the conversation, he explains how WebRTC powers the real-time experiences behind many of today’s most important communication products, and why companies across healthcare, education, customer support, legal, and other regulated industries rely on custom solutions to meet their business and compliance needs. For teams exploring enterprise communication platforms, the episode

Early Voice AI deployments were built on a straightforward pattern: Speech To Text, LLM, Text To Speech. That pipeline was enough to produce compelling prototypes for customer support, sales automation, and meeting summaries. The pattern holds well until it meets a regulated environment. Telecom platforms, telehealth systems, emergency response workflows, and financial infrastructure impose requirements that the basic pipeline was never designed to satisfy.  This post outlines a production voice AI architecture for regulated environments:

Improving quality in WebRTC applications is an ongoing task, it doesn’t stop when you deploy your application. To support maintenance, debugging, and continuous improvement, observability needs to be baked in from the beginning. On March 25, 2026, our guest was Balázs Kreith, a Senior Software Engineer at Riverside.fm, lead developer of the open source ObserveRTC project, and a veteran of WebRTC teams at Whereby and callstats.io. Balázs talks about Quality of Service, Quality of Experience, and share common pitfalls

Over the last few years, Voice AI agents have moved quickly from experimentation into production. Early adoption centered on customer support, basic IVR modernization, sales automation, meeting summaries, and general-purpose voice assistants. These early use cases were low-stakes enough to tolerate imperfection. That is changing. Real-time Voice AI agents are now being deployed in regulated and mission-critical environments such as telecom platforms, telehealth systems, emergency response workflows, and financial infrastructure. In those settings, the hard

Testing an AI voice agent is nothing like testing a standard application. You're validating a live, real-time pipeline where WebRTC audio streaming, speech-to-text, LLM reasoning, and text-to-speech synthesis work together within milliseconds, every time a user speaks. Traditional QA processes and frameworks weren't built for this. They were not designed for systems where latency is a feature, where a 400ms spike in STT processing changes the entire feel of a conversation, or where multi-user dynamics

Recently, I read an article on LinkedIn that captured something many experienced developers have been feeling: software development is changing rapidly in the age of generative AI, but not always in ways we fully understand. One quote especially resonated with me:  “An MIT professor called AI ‘a brand new credit card that lets us accumulate technical debt in ways we were never able to before.’ That credit card now writes 41% of the code.” Whether

When a telehealth platform starts to grow, its video infrastructure either grows with it or becomes the thing that holds it back. That's exactly the crossroads Sessions Health reached. Their HIPAA-compliant practice management platform for mental health professionals had outgrown its self-hosted Jitsi setup, hitting reliability issues at around 500 concurrent sessions. With adoption accelerating, they needed their telehealth video infrastructure to handle 10× that load without compromising call quality or patient privacy. WebRTC.ventures conducted