IoT Mentor

What Embedded / IoT Mentor does

Embedded / IoT Mentor turns an embedded or IoT project brief into a scoped build plan. It can compare platforms such as ESP32, Raspberry Pi Pico, STM32, and nRF52, explain the selection, and name alternatives with the conditions under which they would be preferable.

The planning process covers more than the microcontroller. It separates the hardware purchasing path from the firmware path, checks whether an existing option such as ESPHome, Meshtastic, or Tasmota can avoid custom firmware, and considers how readings will reach a person. Possible outputs include Home Assistant, a device-hosted page, a hosted dashboard, or alerts.

Recommendations are aimed at a working breadboard or MVP. Production and pre-production planning can be introduced when requested, rather than assumed from the beginning.

How it works

Describe the project in ordinary language, including the intended result, previous experience, budget, power source, surroundings, and target timeline. If important constraints remain unclear, the mentor asks a small number of follow-up questions before making a recommendation. It presents the resulting plan in tables where practical, with a short enough format to fit on one screen; further reasoning can be requested.

The analysis estimates time and cost as ranges and identifies the factors behind them. For battery projects, that includes sleep current, running costs, and the effect of deploying multiple nodes. It also distinguishes what a sensor actually measures from what its marketing label may imply, such as inexpensive probes that infer nutrient levels from conductivity.

Setup and configuration

Claude users can download the packaged .skill file from the latest repository release and upload it through Claude's Skills settings. Claude Code users can install it into a user or project skills directory with the included package_skill.py script, install an existing bundle, or unpack the archive manually. Claude Code loads the skill in a later session and lists it through /skills.

A VS Code user can copy the prompt in vscode-copilot/ into Copilot Chat. The repository also provides a Render deployment path for its REST API, and a universal prompt for other AI chat applications. No environment variables are listed for these installation paths.

Tools and capabilities

• Selects a microcontroller, board, and toolchain from project constraints.
• Separates hardware, firmware, connectivity, and user-interface decisions.
• Estimates build time, purchase cost, operating cost, and battery life.
• Identifies power, availability, certification, enclosure, debugging, and learning risks.
• Rejects unsuitable parts or approaches against practical criteria such as poor library maintenance, difficult assembly, or missing debug access.

Limitations and notes

The recommendations stop at an MVP or working breadboard unless a later project phase is requested. The mentor may reject part of a brief when the requested measurement or cost target is not realistic, then continue with a narrower plan. Its examples emphasize practical constraints such as outdoor moisture, radio distance, battery-powered dev boards, sensor accuracy, and hand assembly. It is intended for planning and decision support, not as a substitute for validating hardware, firmware, radio performance, power consumption, or field conditions.