app.py

import gradio as gr
from huggingface_hub import InferenceClient


def transcribe_audio(audio_file_path: str, oauth_token: gr.OAuthToken | None) -> str:
    """Transcribe audio using Inference Providers, billed to the user."""
    if oauth_token is None:
        raise gr.Error("Please sign in with Hugging Face first.")

    client = InferenceClient(provider="auto", token=oauth_token.token)
    transcript = client.automatic_speech_recognition(
        audio=audio_file_path,
        model="openai/whisper-large-v3"
    )
    return transcript.text


def generate_summary(transcript: str, oauth_token: gr.OAuthToken | None) -> str:
    """Generate summary using Inference Providers, billed to the user."""
    if oauth_token is None:
        raise gr.Error("Please sign in with Hugging Face first.")

    client = InferenceClient(provider="auto", token=oauth_token.token)

    prompt = f"""
    Analyze this meeting transcript and provide:
    1. A concise summary of key points
    2. Action items with responsible parties
    3. Important decisions made

    Transcript: {transcript}

    Format with clear sections:
    ## Summary
    ## Action Items
    ## Decisions Made
    """

    response = client.chat.completions.create(
        model="deepseek-ai/DeepSeek-R1-0528:fastest",
        messages=[{"role": "user", "content": prompt}],
        max_tokens=1000
    )
    return response.choices[0].message.content


def process_meeting_audio(audio_file, oauth_token: gr.OAuthToken | None):
    """Process uploaded audio file and return transcript + summary."""
    if oauth_token is None:
        raise gr.Error("Please sign in with Hugging Face first.")

    if audio_file is None:
        raise gr.Error("Please upload an audio file.")

    transcript = transcribe_audio(audio_file, oauth_token)
    summary = generate_summary(transcript, oauth_token)
    return transcript, summary


with gr.Blocks() as app:
    gr.Markdown("# 🎤 AI Meeting Notes")
    gr.Markdown(
        "Sign in with your Hugging Face account, then upload a meeting recording "
        "to get an instant transcript and summary. Inference is billed to your account."
    )

    gr.LoginButton()

    with gr.Row():
        audio_input = gr.Audio(label="Upload Meeting Audio", type="filepath")

    with gr.Row():
        submit_btn = gr.Button("Process", variant="primary")

    with gr.Row():
        transcript_output = gr.Textbox(label="Transcript", lines=10)
        summary_output = gr.Textbox(label="Summary & Action Items", lines=10)

    submit_btn.click(
        fn=process_meeting_audio,
        inputs=[audio_input],
        outputs=[transcript_output, summary_output],
    )


if __name__ == "__main__":
    app.launch()