Here is a little Python script that detects clipboard text and plays it using Kokoro TTS

monitor.py

import os
os.environ["PHONEMIZER_ESPEAK_LIBRARY"] = "C:\\Program Files\\eSpeak NG\\libespeak-ng.dll"
os.environ["PHONEMIZER_ESPEAK_PATH"] = "C:\\Program Files\\eSpeak NG\\espeak-ng.exe"

import sounddevice as sd
import pyperclip
import time
import torch, re
from models import build_model
from kokoro import generate

device = 'cuda' if torch.cuda.is_available() else 'cpu'

MODEL = build_model('pth/kokoro-v0_19.pth', device)
VOICE_NAME = 'af'
VOICEPACK = torch.load(f'voices/{VOICE_NAME}.pt', weights_only=True).to(device)
print(f'Loaded voice: {VOICE_NAME}')

def play_audio_from_text(text):
    sentences = re.split(r'[.\n]+', text)
    for sentence in sentences:
        if sentence.strip() == '':
            continue
        audio, out_ps = generate(MODEL, sentence, VOICEPACK, lang=VOICE_NAME[0])
        sd.play(audio, samplerate=24000)
        sd.wait()

def monitor_clipboard():
    last_clipboard = pyperclip.paste()
    while True:
        time.sleep(0.5)  # Reduce CPU usage
        current_clipboard = pyperclip.paste()
        if current_clipboard and current_clipboard != last_clipboard:
            play_audio_from_text(current_clipboard)
            last_clipboard = current_clipboard

if __name__ == "__main__":
    print("Monitoring clipboard for text changes...")
    monitor_clipboard()

uv.txt

> uv pip install torch --index-url https://download.pytorch.org/whl/cu124
> uv pip install requests
> uv pip install numpy
> uv pip install scipy
> uv pip install phonemizer
> uv pip install munch
> uv pip install transformers
> uv pip install soundfile
> uv pip install pyperclip

And then you can run the code like this:

uv run monitor.py