app.py

import tempfile from pydub import AudioSegment import edge_tts import gradio as gr

language_dict = { "Persian": { "Dilara (Female)": "fa-IR-DilaraNeural", "Farid (Male)": "fa-IR-FaridNeural" } }

async def tts_dialogue_persian(dialogue_text): lines = dialogue_text.strip().split("\n") audio_segments = []

for line in lines:
    if ':' not in line:
        continue
    speaker, text = line.split(":", 1)
    text = text.strip()

    voice = language_dict["Persian"]["Dilara (Female)"] if "زن" in speaker else language_dict["Persian"]["Farid (Male)"]
    communicate = edge_tts.Communicate(text, voice)

    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
        tmp_path = tmp_file.name
        await communicate.save(tmp_path)
        segment = AudioSegment.from_file(tmp_path)
        audio_segments.append(segment)

# Combine all segments
final_audio = sum(audio_segments)
with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
    final_path = tmp_file.name
    final_audio.export(final_path, format="mp3")
return final_path

Gradio interface

with gr.Blocks(title="Persian TTS Dialogue") as demo: gr.HTML("<center><h1>Persian TTS Dialogue (Edge TTS)</h1></center>") gr.Markdown("Use 'زن:' and 'مرد:' as prefixes for lines to select voice.")

with gr.Row():
    with gr.Column():
        input_text = gr.Textbox(lines=10, label="Input Dialogue", placeholder="مرد: سلام\nزن: سلام، خوبی؟")
        run_btn = gr.Button(value="Generate Audio", variant="primary")

    with gr.Column():
        output_audio = gr.Audio(type="filepath", label="Generated Dialogue")

run_btn.click(tts_dialogue_persian, inputs=[input_text], outputs=[output_audio])

if name == "main": demo.queue().launch(share=True)