This is a demo showing the custom GPT I'm developing and posting a series here about. It can easily work with any of your projects in your standard ../Arduino folder. It's multi-platform so it knows where that folder is regardless if you are running Windows, macOS, or Linux.
It talks directly to your board using the arduino-cli tool which is available on all platforms.
It can analyze and edit any of your existing projects all just by talking with it, give you advice about any of them, and compile and upload them all without using any IDE. I'm also posting a series of articles on how to build this and other Customer GPT's using OpenAI. If there is interest I will also develop the same kind of specialized Gemini Gem for Google's AI platform.
Have Fun! (also, check out the GPT's logo 😉)
ripred
Artificial Brains: Arduino and ChatGPT
The Arduino Project Manager GPT isn't a traditional hardware build—it operates at the bleeding edge of IoT integration. This project connects an internet-enabled ESP32 to OpenAI's massive language models, turning a standard microcontroller into an incredibly smart conversational agent.
Securing the API Call
You cannot run AI models on an Arduino. Instead, you use the Cloud.
- The Request: The ESP32 connects to Wi-Fi. It takes a text input (perhaps from a microphone converting speech-to-text, or a serial terminal input).
- The HTTPS POST: The ESP32 crafts a highly secure, encrypted HTTPS request. It attaches your private OpenAI API Key, formats the question using JSON (JavaScript Object Notation), and sends it to
api.openai.com. - The Wait: The Arduino
WiFiClientSecurelibrary waits for the server response. - The Parse: The ESP32 receives the massive JSON response back, uses
ArduinoJSONto cut out the actual answer, and prints it to a screen!
Essential Parts Checklist
- ESP32 or ESP8266 NodeMCU: Standard Arduinos don't have enough RAM or HTTPS capability for this.
- Micro-USB Cable.
- A valid OpenAI API Key.
- A Display: A large TFT or I2C LCD to read the AI's responses.
This architecture forms the backbone of custom "Smart Home Assistant" projects, allowing you to ask your desk lamp complicated hardware questions and receive instant, articulate answers right on an LCD!