AI Voice Assistant Kit — I2S Mic + Amp, Choose Your Board

Everything needed to build a device that hears you, thinks in the cloud, and answers out loud. Digital audio in and out over I2S — no DAC, no analog noise.

What's in every kit

• INMP441 I2S MEMS microphone — digital audio in, no ADC noise
• MAX98357A I2S 3W Class-D amplifier — digital audio out, no DAC needed
• 3W 8Ω speaker
• Tactile button + status LED (wake / mute / state)
• Dupont jumper wire kit
• Wiring diagram and working example firmware

Choose your board

ESP32-DEVKITC (WROOM-32) — the one to pick if you want the device to answer a phone call. It has Bluetooth Classic, so it can pair to a phone as a hands-free headset and sit inside a real call. WiFi for the cloud agent.

ESP32-S3 SuperMini — more RAM and a faster core, better for on-device buffering and richer audio handling. WiFi + BLE. Bluetooth Classic is not available on the S3, so it cannot do the phone-call bridge.

Raspberry Pi Zero 2 W — full Linux. Run Python, use any speech library you like, no firmware flashing. Needs a microSD card and a 5V supply (not included).

⚠ Why there is no Arduino UNO option

An UNO has no WiFi and no I2S peripheral, and 2KB of RAM. It cannot reach a cloud service, cannot drive either of the audio modules in this kit, and cannot buffer speech. We would rather tell you that than sell you a kit you cannot finish.

What you build with it

A voice assistant, an intercom, a reception bot, a talking sensor, or a hands-free call answerer. The microcontroller captures audio and streams it to a server; speech-to-text, your logic, then text-to-speech comes back and plays through the amplifier.