app.py

import os
from flask import Flask, request, jsonify
from flask_cors import CORS
from gpt4all import GPT4All

app = Flask(__name__)
CORS(app)

# Load model once at startup (ensure model file available on the server)
MODEL_NAME = os.environ.get("GPT4ALL_MODEL", "gpt4all-falcon-newbpe-q4_0")
model = GPT4All(MODEL_NAME)

@app.route("/chat", methods=["POST"])
def chat():
    data = request.get_json(force=True)
    prompt = data.get("prompt", "")
    max_tokens = int(data.get("max_tokens", 150))
    # tuned for speed
    with model.chat_session():
        resp = model.generate(
            prompt,
            max_tokens=50,
            temp=float(data.get("temp", 0.1)),
            top_k=int(data.get("top_k", 20)),
            top_p=float(data.get("top_p", 0.7)),
            streaming=False
        )
    return jsonify({"response": resp})

if __name__ == "__main__":
    port = int(os.environ.get("PORT", 5000))
    app.run(host="0.0.0.0", port=port)