"Day 4 – MongoDB Connection & Persistent Storage"

Posted on Thu 16 April 2026 in GenAI

Introduction

In AI systems, data should not be lost after execution. Persistent storage helps store data permanently.

Use

  • Store user inputs
  • Save AI responses
  • Maintain history

On Day 4, I connected my Python application to MongoDB to build a persistent backend system.

Understanding Persistent Storage

Persistent storage means saving data permanently in a database.

Use

  • Prevent data loss
  • Enable data reuse
  • Support long-term storage

Types

  • Stateless → No memory
  • Stateful → Stores memory

MongoDB Overview

MongoDB is a NoSQL database that stores data as JSON-like documents.

Use

  • Store flexible data
  • Handle scalable applications

Structure

  • Database
  • Collection
  • Document

Connecting Python to MongoDB

This step connects Python application with MongoDB.

Use

  • Insert data
  • Retrieve data
  • Build backend systems

Code

from pymongo import MongoClient
from dotenv import load_dotenv
import os
import certifi

load_dotenv()

MONGO_URI = os.getenv("MONGO_URI")
DB_NAME = os.getenv("DB_NAME")

client = MongoClient(MONGO_URI, tls=True, tlsCAFile=certifi.where())

try:
    client.admin.command("ping")
    print("MongoDB Connected ✅")
except Exception as e:
    print("Connection Error ❌", e)

db = client[DB_NAME]
collection = db["ai_outputs"]

Environment Variables

Environment variables store sensitive data securely.

Use

  • Protect credentials
  • Avoid exposing secrets

Code

MONGO_URI=mongodb+srv://<username>:<password>@cluster.mongodb.net/ai_pipeline?retryWrites=true&w=majority
DB_NAME=ai_pipeline

Inserting Data

Inserting data means storing new data into MongoDB.

Use

  • Save AI responses
  • Store user queries

Code

from datetime import datetime, UTC

data = {
    "prompt": "What is AI?",
    "response": "AI is intelligence demonstrated by machines.",
    "timestamp": datetime.now(UTC)
}

result = collection.insert_one(data)
print("Inserted ID:", result.inserted_id)

Retrieving Data

Retrieving data means fetching stored information.

Use

  • Access stored responses
  • View history

Types

Fetch all

for doc in collection.find():
    print(doc)

Fetch one

result = collection.find_one({"prompt": "What is AI?"})
print(result)

Saving AI Output

Saving AI output means storing prompt and response together.

Use

  • Maintain conversation history
  • Enable memory-based AI

Code

def save_ai_output(prompt, response):
    from datetime import datetime, UTC

    data = {
        "prompt": prompt,
        "response": response,
        "timestamp": datetime.now(UTC)
    }

    collection.insert_one(data)

System Flow

System flow shows how data moves in the application.

Use

  • Understand pipeline
  • Debug easily

Flow Diagram

System Flow

-User input -Python processes -AI generates response -Data stored in MongoDB -Data retrieved when needed

Summary

Today, I learned how to connect Python with MongoDB and implement persistent storage.

Key Learning

  • Database integration
  • Data storage and retrieval
  • Building stateful AI systems

This is an important step toward building real-world AI applications.

[GenAI MongoDB Backend]