app.py

import os
from fastapi import FastAPI, Request
from huggingface_hub import InferenceClient

api_key = os.getenv("HF_API_KEY")

client = InferenceClient(
    model="mistralai/Mistral-7B-Instruct-v0.1",
    token=api_key
)

app = FastAPI()

@app.post("/chat")
async def chat(request: Request):
    data = await request.json()
    prompt = data.get("prompt", "")
    response = client.text_generation(prompt=prompt, max_new_tokens=300)
    return {"response": response}