Descriptions:
Joe Reeve, a member of the growth organization at ElevenLabs, presents the technical story behind a viral ‘talk to any statue’ app he built in two hours on a Sunday using Cursor and a single one-shot prompt. The app takes a photo of a statue, runs an OpenAI deep research pass to identify it and generate historical context, calls ElevenLabs’ voice design API to synthesize a period-appropriate voice from a text description, spins up an ElevenLabs conversational agent, and connects the caller — all within roughly 30 seconds.
The talk is grounded in practical detail: Reeve explains the underused ElevenLabs voice design API that accepts a voice description and returns a matching synthetic voice, the agent deployment platform that manages the full phone call lifecycle, and why the end-to-end pipeline can be one-shot in Cursor without traditional software engineering experience. The app went from 50,000 impressions on day one to 1.5 million on day two after a vibe-coding follow-up post went viral, attracting outreach from the Science Museum, Bonhams, Christie’s, and at least one CEO who had spent a year with a ten-person team on a similar problem.
Reeve also discusses what he sees as the next step: embedding the technology physically inside statues with built-in microphones and speakers, removing the screen entirely to avoid the ‘tacked-on technology’ feeling. The talk serves as a practical case study in how ElevenLabs’ voice stack — text-to-speech, voice cloning, and managed agents — can be combined quickly for high-impact interactive experiences.
📺 Source: AI Engineer · Published June 01, 2026
🏷️ Format: Showcase
