CircuitPython 11 adds on-device text-to-speech for microcontrollers
CircuitPython 11 enables real-time text-to-speech synthesis directly on microcontrollers using two open-source speech models selected for different hardware capabilities.
Summary
Adafruit has published a guide for CircuitPython 11's new text-to-speech capability, which runs entirely on microcontroller hardware without external services. The implementation uses Moonshine Klatt for average-capability chips and PicoTTS for more powerful microcontrollers with PSRAM. Previously, microcontroller speech synthesis was considered too slow for practical use.
How the speech models map to hardware
- Moonshine Klatt: selected for average-capability microcontrollers with limited memory
- PicoTTS: used for more powerful microcontrollers equipped with PSRAM
- Performance depends on microcontroller speed and available memory
- No network connectivity required; synthesis runs entirely on-device
Why it matters
On-device speech output expands accessibility options for embedded projects and reduces dependency on cloud services for voice feedback in environments without reliable connectivity.
Original reference: blog.adafruit.com →