Why they stopped
i got the core product to a real mvp, but the next leg stopped being model work and became productization work. the backlog was billing, multi-tenant setup, monitoring, tighter onboarding, and a real distribution path into small businesses. i did not have enough customer pull to justify pushing through that heavier go-to-market and ops layer, so it stalled between working demo and real business.
What it needs
pick one vertical where after-hours calls are painful and repeatable, then sell pilots instead of keeping the product horizontal. dental/medspa, home services, and property management all fit the product shape already. the best next move is a website-to-demo funnel that generates a business-specific setup, then converts that into a paid pilot with basic billing and reliability hardening.
Hard to replicate
the value here is not a landing page or a prompt; it is the full handoff surface that was already shaped around after-hours operations. there is real wiring for live voice calls, knowledge-base retrieval, message capture, appointment intake, call summaries, and a dashboard the owner can check in the morning. there is also a clear product thesis around being the night shift for a business, which is more specific than generic ai receptionist positioning.
About the project
working mvp for an after-hours ai receptionist. the product covers the full loop: twilio + elevenlabs voice handling, fastapi webhook/tooling, a next.js + convex dashboard for calls/messages/appointments, and onboarding that can draft a business knowledge base from a website or uploaded docs. the code is split across separate frontend and backend repos.