Flex-TTS / app.py
FroxAI's picture
Upload 4 files
a5a266c verified
Raw
History Blame Contribute Delete
1.66 kB
import gradio as gr
from src.inference import generate, list_voices
# XTTS-v2's 17 languages with real voice cloning support
LANGUAGES = ["en", "es", "fr", "de", "it", "pt", "pl", "tr", "ru",
"nl", "cs", "ar", "zh-cn", "ja", "hu", "ko", "hi"]
VOICES = list_voices()
def run(text, language, voice):
if not text.strip():
return None, "Enter some text first."
if not voice:
return None, "No voices found — make sure voice_refs/ is included in the repo."
try:
path = generate(text, language, voice)
return path, f"Generated with voice '{voice}' in language '{language}'."
except Exception as e:
return None, f"Error: {e}"
with gr.Blocks(title="FroxAI Flex-Audio") as demo:
gr.Markdown(
"# 🦊 FroxAI — Flex-Audio\n"
"Fine-tuned XTTS-v2 text-to-speech with voice cloning.\n\n"
"⚠️ Inherits the base model's **CPML license (non-commercial use only)**."
)
with gr.Row():
with gr.Column():
text_in = gr.Textbox(label="Text", lines=4, placeholder="Type something to speak...")
language_in = gr.Dropdown(choices=LANGUAGES, value="en", label="Language")
voice_in = gr.Dropdown(choices=VOICES, value=VOICES[0] if VOICES else None, label="Voice")
submit = gr.Button("Generate", variant="primary")
with gr.Column():
audio_out = gr.Audio(label="Output", type="filepath")
status = gr.Textbox(label="Status", interactive=False)
submit.click(run, inputs=[text_in, language_in, voice_in], outputs=[audio_out, status])
if __name__ == "__main__":
demo.launch()