Story
Think about it: voice control is everywhere — in smart TVs, remote controls, smart home devices. And yet, most people stop using it after the first try. Why? Because it only works if you memorize a rigid set of commands. Miss a word, and the system ignores you.
With VoxControl firmware v.2, that experience changes. For the first time, makers can try what future devices should feel like: you speak naturally, and the system reacts to your meaning, not just your words. All of this runs offline, directly on the ESP32-S3 you already have.
Why It Matters
This is more than an upgrade for one board. It’s a glimpse into how tomorrow’s devices will listen and respond. Instead of training yourself to speak like a machine, the machine adapts to you. No cloud services, no internet, no memorized commands — just natural voice in, real actions out. For classrooms, workshops, and demos it makes interaction effortless. And for OEMs watching, it’s a small but real proof-of-concept of where voice control needs to go.
Use Cases
- Basic robot control.
Instead of memorizing one rigid phrase, VoxControl v.2 understands variations. Say “forward, ” “go, ” or “move ahead” — the robot still moves forward. The same works for turns and stopping.
- Classroom experiments.
In a robotics lesson, kids often forget the “official” words. Now one student can say “go left” and another “turn to the left, ” and both will make the robot turn. The focus stays on learning, not on remembering a cheat sheet.
- DIY extensions.
VoxControl isn’t limited to robots. Connect LEDs, a fan, or a small motor, and map “stop / halt” to turning it off, or “go” to switching it on. With firmware v.2, you can extend natural voice control to simple DIY projects instantly.
- Interface prototyping.
If you’re curious how people naturally talk to devices, this firmware is a hands-on way to explore it. By testing different phrases, you can see which ones feel intuitive and how a recognition system should respond.
How It Works
Under the hood, it’s a neat trick. The microphone picks up speech, an ASR model turns it into text, and a classifier maps different phrases into the same action. “Forward, ”“go ahead, ” or “move on” all trigger the same behavior. What looks like understanding is actually a carefully optimized recognition pipeline, packed into the ESP32-S3 and running entirely at the edge.
Try It Yourself
Owners of VoxControl can request firmware v.2, flash it, and see the change instantly. Say commands the way you want, and the board responds. Moving forward, turning, stopping — it all feels smoother and more natural when you don’t have to remember the “correct” phrase.
Feel free to message us on Hackster or email us at hi@grovety.com to request the firmware update.
What’s Next
This update is only the beginning. If many of you ask for the same new words or actions, we’ll include them in future firmware releases. By trying VoxControl v.2, you’re not just upgrading a board — you’re helping shape how natural voice control will work in the future.
Share Your Feedback
What words worked best for you? Which ones do you wish were included? Your input will guide the next round of updates and bring us all closer to devices that listen the way we expect them to.
Links
