app.py

from fastapi import FastAPI
from pydantic import BaseModel
from huggingface_hub import InferenceClient
import os

app = FastAPI()

client = InferenceClient(
    api_key=os.getenv("HF_TOKEN")
)

class ChatRequest(BaseModel):
    model: str
    messages: list

@app.get("/v1/models")
def models():
    return {
        "data": [
            {
                "id": "Qwen/Qwen2.5-Coder-14B-Instruct",
                "object": "model"
            }
        ]
    }

@app.post("/v1/chat/completions")
async def chat(req: ChatRequest):

    response = client.chat.completions.create(
        model=req.model,
        messages=req.messages,
        max_tokens=2048
    )

    return response