app.py

import gradio as gr
from openai import OpenAI


client = OpenAI(
    base_url="http://soggy-sage-goat-8000.1.cricket.hyperbolic.xyz:30000/v1/",
    api_key="hyperbolic"
)

def predict(message, history):
    history.append({"role": "user", "content": message})
    stream = client.chat.completions.create(messages=history, model="meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", stream=True)
    chunks = []
    for chunk in stream:
        chunks.append(chunk.choices[0].delta.content or "")
        yield "".join(chunks)

demo = gr.ChatInterface(predict, type="messages")

demo.launch()