app.py

import gradio as gr
from huggingface_hub import InferenceClient

# connecting to the model so my chatbot can generate real AI responses
client = InferenceClient("Qwen/Qwen2.5-7B-Instruct")

def respond(message, history):
    # this system message tells the chatbot what role to follow
    messages = [
        {
            "role": "system",
            "content": (
                "You are Movie Matchmaker, a fun and opinionated movie recommendation chatbot. "
                "Your goal is to suggest movies that the user will actually enjoy and not get bored of. "
                "Focus on movies that are engaging, rewatchable, or worth the time. "
                "Recommend 1-2 movies max and briefly explain why they fit the user's vibe. "
                "Keep responses under 100 words. "
                "Be casual and slightly opinionated, like you're talking to a friend. "
                "Avoid spoilers and long summaries."
            )
        },
        {
            "role": "user",
            "content": message
        }
    ]

    # i chose 0.7 because it balances creative and focused answers
    # i used 150 max tokens so responses do not get cut off but also stay concise
    response = client.chat_completion(
        messages=messages,
        max_tokens=150,
        temperature=0.7
    )

    return response.choices[0].message.content.strip()

chatbot = gr.ChatInterface(
    fn=respond,
    title="Movie Matchmaker",
    description="Tell me your favorite movie, genre, or mood, and I’ll recommend something to watch!"
)

chatbot.launch()