Multiplayer Conversational voice agent
Team consisting of Qordinate Co-founders building AI Coordination OS, skilled in Multi-Agent Systems, RAG, LLM Observability, Python, and AWS Lambda.
YouTube Video
Project Description
The World’s First Multiplayer Voice Agent with Real-World Agency.
The Problem
Voice AI has historically been a lonely experience - a 1-on-1 silo between a user and a bot. But real work happens in teams. When a production incident strikes or a complex customer issue arises, you don’t want a chatbot; you want a teammate who can join the conference call, listen to the context, and execute tasks without you having to open your laptop.
The Solution
John is a multi-agent orchestration platform that treats AI as a first-class participant in your group calls. John sits in the conference line as a silent observer, processing the conversation in real-time. It understands who is speaking and ensures strict authorization protocols are met.
Key Innovation: Voice-to-Action via MCP
We didn’t just build a talkative bot; we built a doer. Using the Model Context Protocol (MCP), John bridges the gap between telephony and your dev tools.
- Context Aware: John listens to the human discussion to understand the root cause.
- Secure: John knows that only “Rahul” has the permission to request a deploy.
- Agentic: In our demo, John connects to a Cursor agent to spin up a new git branch and write code to fix a health-check failure—all triggered by a simple voice command during a phone call.
Team
Products & Tools
Additional Links
GitHub repo