Building a Multi-Agent Voice Command Center: Load Bearing Empire's QHV Stack on Supabase, VAPI, and Vercel

Description and use

A voice command center with 8 AI agents and 8 workflows, built on Supabase Edge Functions, VAPI, and Vercel, delivering real‑time state consistency across interfaces.

Problem

Managing multiple AI agents and automation workflows for voice‑driven operations while ensuring real‑time state consistency across several front‑end interfaces.

Outcome

The architecture enabled reliable, concurrent voice‑driven workflows with consistent state across multiple interfaces, moving the service from proof‑of‑concept to production.

Workflow

• Design an 8‑table Postgres schema to coordinate operations and a single webhook listener
• Implement VAPI agent callbacks that trigger state mutations
• Deploy Supabase Edge Functions to process callbacks and update state
• Use Vercel for front‑end interfaces that consume real‑time state
• Integrate with a self‑hosted Asterisk PBX backbone for voice routing

Evidence

We built a production voice command center handling 8 concurrent AI agents, 8 automation workflows, and real‑time state management across 5 frontend interfaces—all on Supabase Edge Functions, VAPI's voice API, and Vercel deployments.

Source

how we designed an 8-table Postgres schema to coordinate valet trash operations, real estate wholesaling, and credit repair workflows through a single webhook listener; how we structured VAPI agent callbacks to trigger state mutations without race conditions; and the specific infrastructure decisions that let us ship this from a self-hosted Asterisk PBX backbone.

Source