Voice AI · Case study

CallFlow

Client: A multi-location services businessDiscipline: Voice AIStatus: Live demo

Real-time voice agents that book appointments and answer calls, right in the browser.

The problem

Front desks miss calls, and callers hate phone trees. Our client wanted natural voice agents that could hold a real conversation and complete tasks, without locking into an expensive voice platform.

Our approach

We built real-time voice agents directly on the OpenAI Realtime API, owning the WebRTC layer ourselves to keep latency low and avoid vendor lock-in. Each agent runs a distinct business scenario and can be wired into CRMs, calendars, and knowledge bases through tool calls.

What we shipped

Three live voice agents you can talk to: appointment booking, lead qualification, and support. Sub-second responses, with clean call handling across browsers and mobile.

  • OpenAI Realtime voice agents
  • Direct WebRTC, no lock-in
  • Multiple business scenarios
  • Cross-browser, mobile-hardened
Next.jsOpenAI Realtime APIWebRTCUpstash

What we took away

Going direct to the API instead of a voice platform meant more code, but far lower latency and cost, plus the freedom to shape the conversation exactly how the business needed.

See CallFlow in action

This is a real, working build. Open it and try it yourself.

Visit the live demo
Want something like this?

Let's build your version.

Tell us the problem. We'll tell you how we'd ship it.