app.py

import gradio as gr
from elevenlabs.client import ElevenLabs
import tempfile
import os

def generate_speech(text = "", api_key = "", voice_id = "JBFqnCBsd6RMkjVDRZzb", model_id = "eleven_v3", language_code = "en", output_format = "mp3_22050_32"):
    """
    Convert input text to speech using ElevenLabs API and return the audio file path.

    Args:
    text (str): Required. The input text to be converted into speech.
    api_key (str): Required. The ElevenLabs API key used to authenticate the request. This should be kept secret and treated like a password.

    voice_id (str): The ID of the voice to use for speech generation. Default: "JBFqnCBsd6RMkjVDRZzb". Other available voices can be found at: https://elevenlabs.io/app/default-voices

    model_id (str): The ID of the ElevenLabs model to use. Typically one of: "eleven_v3" (default) or "eleven_multilingual_v2"

    language_code (str): The two-letter ISO 639-1 language code specifying the language of the input text. Example: "en" for English, "lv" for Latvian. Full list: https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes

    output_format (str): The desired output audio format. Default: "mp3_22050_32". Other formats are listed here: https://elevenlabs.io/docs/api-reference/text-to-speech/convert#request.query.output_format.output_format   

    Returns:
    tuple
        A tuple containing:
    
        audio_file_path : str or None
            The file path to the generated audio file if the speech synthesis was successful, or None if an error occurred.
        
        status_message : str
            A human-readable message indicating the result of the operation, such as a success confirmation or an error description.
    """
    if not text.strip():
        return None, "❌ Text is empty."
    if not api_key.strip():
        return None, "❌ API key is required."

    try:
        # Initialize ElevenLabs client dynamically with the provided key
        elevenlabs = ElevenLabs(api_key=api_key)

        # Create a temporary file to store the generated speech
        temp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3")

        # Convert text to speech in chunks
        with open(temp_file.name, "wb") as f:
            for chunk in elevenlabs.text_to_speech.convert(
                voice_id=voice_id,
                output_format=output_format,
                text=text,
                language_code=language_code,
                model_id=model_id
            ):
                if chunk:
                    f.write(chunk)

        return temp_file.name, "✅ Speech generated successfully."

    except Exception as e:
        return None, f"❌ Error: {str(e)}"


with gr.Blocks(title="ElevenLabs Text-to-Speech") as demo:
    gr.Markdown("## 🗣️ ElevenLabs Text-to-Speech Generator")
    gr.Markdown(
        """
        Convert your text into speech using the [ElevenLabs API](https://elevenlabs.io/).  
        Fill in the fields below and click **Generate Speech**.
        """
    )

    with gr.Row():
        api_key = gr.Textbox(
            label="🔑 ElevenLabs API Key",
            type="password",
            placeholder="Enter your ElevenLabs API key",
        )

    with gr.Row():
        text_input = gr.Textbox(
            label="📝 Text to Convert",
            placeholder="Enter the text you want to convert to speech...",
            lines=8,
        )

    with gr.Row():
        voice_id = gr.Textbox(
            label="🎤 Voice ID",
            value="JBFqnCBsd6RMkjVDRZzb",
            info="Default: JBFqnCBsd6RMkjVDRZzb (Rachel). Other voices: https://elevenlabs.io/app/default-voices",
        )

    with gr.Row():
        model_id = gr.Dropdown(
            ["eleven_v3", "eleven_multilingual_v2"],
            value="eleven_v3",
            label="🧠 Model",
            info="Choose ElevenLabs model (default: eleven_v3)",
        )

    with gr.Row():
        language_code = gr.Textbox(
            label="🌐 Language Code",
            value="en",
            info="Language code must follow ISO 639-1 (e.g., en, lv, fr). See: https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes",
        )

    with gr.Row():
        output_format = gr.Textbox(
            label="📁 Output Format",
            value="mp3_22050_32",
            info="Default: mp3_22050_32. Other formats: https://elevenlabs.io/docs/api-reference/text-to-speech/convert#request.query.output_format.output_format",
        )

    with gr.Row():
        generate_btn = gr.Button("🚀 Generate Speech")

    with gr.Row():
        audio_output = gr.Audio(label="🔊 Generated Audio", type="filepath")
        status_msg = gr.Markdown()

    generate_btn.click(
        fn=generate_speech,
        inputs=[text_input, api_key, voice_id, model_id, language_code, output_format],
        outputs=[audio_output, status_msg],
    )


if __name__ == "__main__":
    demo.launch(mcp_server=True)