The Browser Agent SDK is live: four composable npm packages that drop a Deepgram voice agent into any web app.
Widget → React components → React hooks → framework-agnostic core. Install the layer you want, the rest comes with it.
Reconnection, audio buffering, playback-aware mode tracking, KeepAlive, Silero VAD, typed events: same across every layer.
The hard parts of browser voice are handled once and inherited everywhere.
Full details, six layouts, and the path from widget to vanilla JS in the blog post:
https://t.co/TphMtNngUe
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
Why it matters
The browser half of a voice AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition →, reconnection, audio buffering, VAD, and tracking when the agent is speaking, is where teams lose a sprint. This handles those once and lets you choose the abstraction level rather than rebuilding at each.
