Home assistant use Hassil for parsing the intent, but it is not suitable for my current use case. Cause I want to use my native Vietnamese, as well as the microphone quality is not good enough for far distance so I wrote a parser myself with the given intent and provided the output command.
I am using ESP32 with INMP441 mic, as well as a Raspberry Pi 4 as a voice satellite. For the Esp32, I can create a text sensor and update it whenever a STT is completed. For the Pi 4, I have to create a MQTT sensor.
Could voice assistant expose a text sensor itself for a “last command”, it would be easier for people like me to control the intent ourself