Voice DSP+ Core Board
The voice engine
for your next build.
The Voice DSP+ board for your Raspberry Pi, robot, custom enclosure or do-it-yourself project. Four microphones, a dedicated XVF3800 voice processor and powerful audio outputs, without an enclosure or speakers you do not need.
Choose your host, choose your speakers and put voice interaction where you want it.
Backorders are now closed. A new batch is tentatively expected around 20 November 2026 (estimated, not guaranteed). Join the waiting list. This is the new board-only configuration. Enclosure, speakers, Raspberry Pi and CoreZero S3 are not included.
01 / Raspberry Pi
Your project. Full control.
Fit the board to a standard 40-pin Raspberry Pi header for I2S audio, I2C control and SPI integration. Build a voice-enabled robot, intercom, kiosk or local assistant without developing the entire microphone front end yourself.
For 40-pin Raspberry Pi models including Zero / Zero 2 W, 4 and 5. A fitted header and model-specific software are required.
02 / USB sound card
Plug in. Select. Talk.
With compatible USB firmware, use the board as a standard processed microphone and playback device for your computer. Ideal for voice applications, dictation, conferencing and rapid experiments.
Use firmware for the exact board revision. V1 firmware is not a V2 update; the final V2 software integration is in progress.
Built-in features
Capture. Process. Play.
- Four integrated digital microphones in a square array.
- Dedicated XVF3800 voice DSP: beamforming, acoustic echo cancellation, noise suppression, dereverberation and automatic gain control.
- Powerful stereo amplifier: 2 × 25 W capability with suitable 20 V power and 4-ohm speakers.
- Separate 2 × 2 W stereo outputs for small voice speakers. Speakers are not included.
- 3.5 mm stereo output for a suitable powered speaker or stereo system, with connection detection and firmware-controlled routing.
- 24 addressable RGB LEDs for voice direction, status and custom feedback.
- Four physical buttons with configurable actions.
- PAJ7620U2 gesture sensor for touch-free controls.
- 40-pin Raspberry Pi interface plus USB audio operation.
- External microphone-array connection and 3.3 V Qwiic I2C expansion.
Wake words and AI run on your controller or host, not on the XVF3800. Control mappings and audio routing depend on the installed firmware. Maximum amplifier capability is not available from a normal computer USB port, and the complete power and thermal budget limits continuous output.
Optional microphone extensions
Move the microphones.
Keep the voice engine.
Four microphones are already on the main board. Choose an external array when your enclosure needs a different microphone position or geometry. Each extension includes a straight 20 cm shielded ribbon cable.
LINE array
Four microphones for front-facing robots, displays and intercom panels. PCB approximately 110 × 11 mm.
SQUARE array
Four microphones for central tabletop placement and all-around applications. PCB approximately 54 × 54 mm.
Optional accessories, sold separately and available to preorder. Use the internal array or one external array at a time, with matching firmware. Shielding helps limit interference; it is not a guarantee of electromagnetic immunity.
In the box
Voice DSP+ Core Board with four integrated microphones, dedicated audio DSP, amplifiers, RGB LEDs, buttons and gesture sensor.
No enclosure, speaker, Raspberry Pi, external microphone array, power supply or ESP32-S3 controller included.
Add your own controller
Prefer a Wi-Fi voice satellite? Add the optional CoreZero S3 ESP32-S3 for ESPHome and Home Assistant integration. It is not included in the bare-board pack.
For a packaged device with CoreZero S3 and small speakers already inside, choose Compact Voice. For the assembled two-way music and voice system, choose Full Speaker.
Documentation and firmware
Follow the guide for your exact hardware revision and host. The board is an audio front end, not a complete assistant server. Your software supplies wake-word detection, speech recognition, answers and speech synthesis.
Before you build
Does the board include a speaker or CoreZero S3?
No. Both are optional for the Core Board. Compact Voice and Full Speaker include CoreZero S3 and speakers.
Can I use it without a Raspberry Pi?
Yes, as a USB sound card with compatible firmware. You can also add an optional ESP32-S3 controller for a voice-satellite build.
Can I connect passive speakers?
Yes, suitable passive speakers connect to the amplifier terminals. For powered speakers or a stereo system, use the stereo jack. Match power, impedance and wiring to the integration guide.
Is this available for immediate delivery?
No. Backorders are now closed. The next batch is tentatively expected around 20 November 2026, subject to production. Joining the waiting list is free and does not place an order or reserve a unit. Existing orders are not changed by this notice.
