Real-time Persian speech recognition. Choose a model, press 'Record Audio', and start speaking.
The 'Large' model is more accurate but slower. The 'Small' model is faster.