Physical Hermes: app and UI for Hermes Agent

Turn a cheap ESP32-S3 AI robot toy into a voice device for your own Hermes Agent

What Physical Hermes does

You flash prebuilt firmware to the robot with esptool, give it your home Wi-Fi through a setup page on your phone, then connect it to Hermes with the hermes-gadget-sdk plugin. Hold the boot button, talk, and Hermes answers out loud, while the round screen acts as Hermes's face and shows replies. Swiping down cancels a reply, and the battery level appears on screen.

The robot behaves as a real Hermes client, so tools, memory, skills and cron all work. Speech recognition and text-to-speech use whatever is configured in Hermes. Your Wi-Fi password goes from your phone to the robot without passing through a server.

The README names the robot (about AU$36) and notes that a firewall rule for port 8765 is the step most people miss. It also describes porting the work to a different robot.

Key features

Prebuilt firmware flashed with esptool
Wi-Fi setup from a phone, with no keyboard on the robot
Push-to-talk using the boot button
Replies shown on screen and spoken aloud
Battery-powered and portable
Uses the Hermes speech settings

When to use it

Talking to your Hermes Agent without a keyboard or phone
Building a desk companion that runs on your own model
Porting the firmware to a different ESP32 board

Who it is for: Hermes Agent users comfortable flashing hardware who want a physical voice interface.

How it fits with Hermes Agent

It is built for Hermes Agent and connects through the hermes-gadget-sdk plugin and a gadget platform in the gateway.

How to install Physical Hermes

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

hermes plugins install https://github.com/Adolanium/hermes-gadget-sdk/tree/main/plugin --enable
hermes config set platforms.gadget.enabled true
hermes gateway restart
hermes gadget info

Requirements: ESP32-S3 1.28" AI Voice Chat Robot, a USB-C data cable, a computer to flash it, Hermes Agent on the same network, port 8765 reachable, speech-to-text and text-to-speech configured, ffmpeg if speech output is not WAV

Note: It requires buying a specific ESP32-S3 robot and flashing its firmware.

FAQ

What is Physical Hermes?

Physical Hermes is a project that turns an inexpensive ESP32-S3 AI robot toy into a voice device for Hermes Agent. It uses your own machine and model.

Does Physical Hermes work with Hermes Agent?

Yes, it is made for it. The robot connects as a Hermes client, so tools, memory, skills and cron are available.

What do I need to run Physical Hermes?

You need the specific ESP32-S3 robot, a USB-C data cable, a computer for flashing, and a working Hermes Agent on the same network. Port 8765 must be reachable from your LAN.

Is Physical Hermes free and open source?

The license field is not recognized (NOASSERTION), so check the repository's license text or ask the author before reuse.