ForestHub Logo ForestHub Logo ForestHub Tools

Board Compatibility Checker

Search for your dev board — see every AI use case it can handle.

Board Compatibility Checker

Find what your dev board can run

516 AI-capable boards indexed

How the Board Checker Works

  1. Search for your board — Type the name of your dev board. We have 516 AI-capable boards from ESP32, STM32, nRF52, RP2040, and more platforms.
  2. See the full compatibility matrix — Every use case is scored against your board's actual RAM, flash, and clock speed, combined with its MCU family's architecture features.
  3. Explore detailed analyses — Expand any use case to see technical reasoning, best framework recommendation, and links to in-depth guides.

Need help choosing hardware? Try the MCU Compatibility Checker or compare AI frameworks side-by-side.

Frequently Asked Questions

How do I know if my dev board supports edge AI?
A board needs sufficient RAM (64 KB minimum for simple models, 256+ KB for vision tasks), flash storage for the model binary, and compatible architecture. This tool checks all requirements automatically against verified specs for 516 AI-capable boards.
What is the difference between board specs and MCU family specs?
An MCU family defines the processor architecture, instruction set (FPU, DSP, SIMD), and peripheral interfaces. A specific dev board may have different RAM and flash amounts depending on the module used. This tool uses your board's actual specs for accurate scoring.
Can I run object detection on an ESP32 dev board?
ESP32-S3 boards with camera support (like the ESP32-S3-EYE or Freenove ESP32-S3-WROOM) can run quantized object detection models via TensorFlow Lite Micro. Standard ESP32 boards lack the vector instructions and RAM headroom needed for reliable vision inference.
Where does the board data come from?
Board specifications are synced from the PlatformIO registry and cross-referenced with manufacturer datasheets. Each board is mapped to its MCU family for architecture-level compatibility checks. Data is updated regularly.