JARVIS Desk Bot: a talking ESP32 desk robot for your AI agent

JARVIS Desk Bot chains an ESP32-S3 desk robot to xiaozhi-esp32-server (local Parakeet speech recognition and Kokoro text-to-speech) to any Hermes agent. It plays a wake chime, shows 'thinking' eyes, speaks short status lines so there's no dead air, and reports back when a background helper finishes. Demos include Google Tasks. It went viral on X this week (via @witcheer and the maker @vineethkalluru).

Hardware & cost: An ESP32-S3 desk robot running Espressif's ESP-Brookesia chatbot example with the xiaozhi agent (release/v0.7, ESP-IDF 5.5), a machine for the voice server, and a Hermes agent. ESP32-S3 voice boards typically cost tens of dollars. MIT licensed.

Difficulty: Medium. Flashing ESP-IDF firmware and running a local speech server.

Agent install prompt (paste into Claude Code, Cursor or Codex): "Clone https://github.com/dreamhunter02/jarvis-desk-bot. Follow its README: set up xiaozhi-esp32-server with Parakeet ASR and Kokoro TTS, run the repo's apply.py installer to add the Hermes provider and chimes, flash my ESP32-S3 robot with the ESP-Brookesia chatbot example, and point it at my server. Ask before flashing."

Seen on X: https://x.com/witcheer/status/2111028813555523855 and the launch post https://x.com/vineethkalluru/status/2108388089731436546

Source: dreamhunter02/jarvis-desk-bot on GitHub — https://github.com/dreamhunter02/jarvis-desk-bot (facts checked against the project README, Oct 2026; prices vary by region).