Skip to content
Taveras Solutions · Go to home
Book a diagnostic

MITESP32-S3

esp32-data-panel

A physical ESP32-S3 panel that queries the Instagram API on its own.

A 2.8" desk screen that shows followers, reach, and recent reels with no PC, no MQTT, and no Home Assistant. It includes a Claude Code skill that builds the panel with you, with this or another data source.

Stars
1
Forks
0
Language
ESP32-S3
License
MIT

THE PROBLEM

A panel that talks to an API on its own means solving HTTPS, display, touch, and audio on a microcontroller, and every detail costs an afternoon. The repo documents those traps and leaves the firmware working.

Who it's for

Creators who want their metrics in view, and makers who want a proven starting point for ESP32-S3 panels.

WHAT IT DOES

  • Queries Meta's Graph API over verified HTTPS, straight from the board.
  • Six swipeable views: live, followers, latest reels, reach, growth, and system.
  • A double-tap voice assistant that talks to OpenAI's audio API.
  • A probe firmware that characterizes the board: pull-ups and I2C scanning.
  • A second firmware that monitors Claude Code spend from a LAN service.
  • A Claude Code skill with the pinout, the bring-up ladder, and the traps, so the agent builds the panel with you.

How it's built

  1. Probe probe/ detects which GPIOs have pull-ups and scans I2C before flashing the real firmware.
  2. Data ig-panel-fw calls the Graph API with pinned root certificates and tells zero apart from unreported data.
  3. Interface LovyanGFX and LVGL draw a tileview with swipe, long press, and double tap.
  4. Voice asistente.cpp records through the ES8311 codec and sends the audio to OpenAI to play back the answer.
  5. Skill skill/ packages hardware, design, data, and voice so Claude Code can repeat the build.

Stack

ESP32-S3C++CInstagram APIMeta APIOpenAIClaudePython

Why we built it

It packs everything learned while building it into a skill, so the next panel does not repeat the same lost afternoons. It shows how Taveras Solutions connects hardware, APIs, and AI end to end.

Getting started

pio run -d probe -t upload && python3 monitor.py
cp ig-panel-fw/src/config.example.h ig-panel-fw/src/config.h  # and fill in the secrets
pio run -d ig-panel-fw -t upload && python3 monitor.py
Shortcut: cp -R skill ~/.claude/skills/esp32-data-panel

Need something like this built for you?

We use these same tools in our client projects.