How to Intermediate · 3 min read

How to give AI agent database access

Quick answer

To give an AI agent database access, implement a secure API or direct database connection that the agent can query dynamically. Use code to connect the agent to the database via libraries like psycopg2 for PostgreSQL or sqlite3, and ensure queries are sanitized to prevent injection.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0
Database access credentials
pip install psycopg2-binary or sqlite3 (built-in)

Setup

Install necessary Python packages and set environment variables for API keys and database credentials.

Use pip install openai psycopg2-binary for PostgreSQL or sqlite3 is built-in.
Set environment variables for OPENAI_API_KEY and database connection info.

bash

pip install openai psycopg2-binary

Step by step

This example shows how to connect an AI agent to a PostgreSQL database, run a query, and use the result in an AI prompt.

python

import os
import psycopg2
from openai import OpenAI

# Connect to PostgreSQL database
conn = psycopg2.connect(
    dbname=os.environ['DB_NAME'],
    user=os.environ['DB_USER'],
    password=os.environ['DB_PASSWORD'],
    host=os.environ['DB_HOST'],
    port=os.environ.get('DB_PORT', 5432)
)
cur = conn.cursor()

# Query example
cur.execute("SELECT name, price FROM products WHERE id = %s", (1,))
product = cur.fetchone()

# Prepare prompt with database data
prompt = f"Product info: Name={product[0]}, Price=${product[1]}. Suggest a marketing tagline."

# Initialize OpenAI client
client = OpenAI(api_key=os.environ['OPENAI_API_KEY'])

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": prompt}]
)

print(response.choices[0].message.content)

# Cleanup
cur.close()
conn.close()

output

Innovative and affordable, the ProductName is your perfect choice for quality and value!

Common variations

You can adapt this pattern for other databases like SQLite or MySQL by changing the connection library and query syntax. Async database clients and streaming AI responses are also possible.

python

import os
import sqlite3
from openai import OpenAI

# SQLite connection
conn = sqlite3.connect('example.db')
cur = conn.cursor()

cur.execute("SELECT name, price FROM products WHERE id = ?", (1,))
product = cur.fetchone()

prompt = f"Product info: Name={product[0]}, Price=${product[1]}. Suggest a marketing tagline."

client = OpenAI(api_key=os.environ['OPENAI_API_KEY'])
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": prompt}]
)

print(response.choices[0].message.content)

cur.close()
conn.close()

output

Discover the unbeatable value of ProductName — quality meets affordability!

Troubleshooting

If you get connection errors, verify your database credentials and network access.
For query errors, check SQL syntax and parameterization to avoid injection.
If AI responses are irrelevant, ensure the prompt includes clear context from the database.

✅

Key Takeaways

Use parameterized queries to securely fetch data for AI prompts.
Connect AI agents to databases via APIs or direct DB clients depending on architecture.
Always sanitize inputs and handle credentials securely using environment variables.

Verified 2026-04 · gpt-4o

Verify ↗