The Arrival of ESP32-Based Voice Hardware

Alasdair Allan
3 min readJan 24, 2018

Amazon’s $100 million Alexa Fund has paid dividends for them with a stealth takeover of CES for the last couple of years, where we’ve seeing third-party manufacturers integrating Amazon’s technology into devices ranging from cameras to cars. But this year’s take over of CES was anything but stealthy, because Amazon Alexa and Google Assistant went head-to-head in Vegas, and arguably Google came out ahead. While the donuts might have helped, it’s clear that Google has put a lot of effort into appealing to developers, and catching up with Amazon’s apparent lead.

Although both Amazon and Google have shipped their own development kits, the two big players have very different approaches. Amazon’s hardware is aimed squarely at commercial manufacturers, while Google’s kits are aimed at makers and hobbyists rather than companies.

However over the last year, we’ve also seen a big push by third-party manufacturers to provide development hardware; unfortunately this has tended to be expensive, with the odd exception.

Espressif ESP32 LyraTD MS1 HDK (📷: Espressif)

So it’s rather interesting to suddenly see a number of third-party hardware development kits based around the ESP32.

Espressif Systems is bringing out its own kit, the ESP32 LyraTD MS1 HDK, which had the softest launch any piece of hardware could ever have being announced rather stealthily on Twitter with the question, “does anyone need something like that?” the answer to which seemed to be “Yes, when can I have it?”

Based around the ESP32, the board comes with WiFi and BLE, and has a four-microphone array, and dual speaker output ports.

The board also supports a micro SD card for audio files, and breaks out pins for expansion including I2C, SPI, UART, and I2S.

Interestingly, the kit supports all the major voice services; Amazon Alexa, Google Assistant, and Baidu DuerOS, with hot word recognition running locally on ESP32 processor.

While the shipping date is tentatively “next week,” it’s not yet clear where they will be initially available, and we’re not yet certain what the price point will be. The announcement from Espressif also had the somewhat surprising effect of surfacing more voice boards based around their ESP32.

The ESP32 Circle from WhyEngineer. (📷: Baoshi)

The ESP32 Circle is built by WhyEngineer, and has a single microphone. But if you can’t wait for the official board from Espressif Systems, it has the advantage of being available right now on Taobao for ¥169 (that’s just $26). More information on the ESP32 Circle can be found on their Wiki.

It’s interesting to see the rapid spiral downward in price of voice hardware development kits, because while the cheaper options might not support the far-field microphones of the more expensive kits, it’s possible they’ll be just good enough.