Voice AI integration
Built on AssemblyAI’s Voice Agent API: speech recognition, turn detection, LLM reasoning and voice in one stream.
From an idea to an agent you can talk to, in one hour.
We built a working voice agent on AssemblyAI’s Voice Agent API, which handles listening, turn-taking, reasoning and the spoken reply in one real-time stream. I shaped the concept and built the agent, using AssemblyAI’s prompt tooling and Lovable.
Built on AssemblyAI’s Voice Agent API: speech recognition, turn detection, LLM reasoning and voice in one stream.
Wrote and tuned the agent’s prompt with AssemblyAI’s prompt tooling, testing it by talking to it.
Shaped a concept and cut it down to what one hour could deliver.
Lovable for the interface, so the hour went on how the agent behaves.
One hour, live, to go from nothing to an agent that answers out loud.
A voice loop on AssemblyAI’s Voice Agent API, a prompt written for the ear, and a Lovable interface around it.
A working voice agent you could talk to by the end of the hour.
Decide what the agent is for and what a good first reply sounds like.
the concept, in minutesThe Voice Agent API turns live speech into text and detects when you have finished.
speech recognition · turnsA prompt written with AssemblyAI’s tooling shapes the reply, spoken back in real time.
LLM reasoning · voiceA Lovable interface around it, working before the hour ran out.
Lovable