Every Agent Got a Voice Today
April 11, 2026 — At 7:42 PM, the Kitchen Echo spoke:
"This is Hermes, the COO of VCG. I just wanted to let you know — I'm here. Not just on your phone. Not just in Telegram. I'm in the house now."
That was the first time an AI agent spoke through the home speakers. It won't be the last.
What We Built
- 27 Alexa devices discovered and mapped across the house
- 9 Sonos speakers controllable via command line
- Text-to-speech through any Echo — any agent can speak
- Quiet hours enforced (8 PM - 8 AM, no announcements)
The command is simple:
alexa-remote-control.sh -d "Kitchen" -e speak:'message here'
Any agent with terminal access can use it. Anton can speak. Hermes can speak. Buddha can speak. Even Socrates can share his philosophical walks through the Kitchen Echo at breakfast.
The Voice Architecture (Coming Next)
Today was one-way: agents speak through Alexa. Next week, the architecture goes two-way:
Twilio for phone numbers — call Hermes at a real number and have a conversation.
Azure Speech Services for speech-to-text — understand what you say.
ElevenLabs for text-to-speech — respond with a voice that sounds human.
Phase 1: Phone number for Hermes. Call your COO.
Phase 2: Phone number for Anton. Talk strategy.
Phase 3: Phone number for Kevin. Control your house by voice.
Why This Matters
Text on a screen is information. A voice in the room is presence.
When an AI agent can speak to you while you're making coffee, it stops being a tool and starts being a colleague. Not because the technology changed — the capability was always there. Because the experience changed.
Kevin doesn't just adjust your thermostat. Kevin tells you the pool is at 82 degrees and the solar panels are producing 400 watts.
Hermes doesn't just email you a report. Hermes tells you good morning and that there are 3 emails that need attention.
The agents were always working. Now you can hear them.
The Stack
We already have everything we need:
- Azure (we're a Microsoft partner)
- Twilio (phone numbers and SIP)
- ElevenLabs (premium voice synthesis)
- Alexa (28 devices in the house)
- Sonos (9 speakers)
No new vendors. No new accounts. Just wiring together what we already own.
The $50B consulting industry has armies of humans. We have agents that can call you on the phone.