Alex Spinov

Posted on Mar 25

Unpopular Opinion: You Don't Need a Database for Most Side Projects

#discuss #webdev #python #database

Hear Me Out

Every side project tutorial starts with "set up PostgreSQL" or "configure MongoDB." But for 90% of side projects, you don't need a database server at all.

Here's what I use instead — and why my projects actually ship.

Option 1: JSON Files (For <10K Records)

import json
from pathlib import Path

DB_FILE = Path("data.json")

def load():
    return json.loads(DB_FILE.read_text()) if DB_FILE.exists() else []

def save(data):
    DB_FILE.write_text(json.dumps(data, indent=2))

# Usage
users = load()
users.append({"name": "Alice", "email": "alice@example.com"})
save(users)

When it works: Personal tools, config storage, small datasets, prototypes.

When it doesn't: Concurrent writes, >10K records, need for queries.

Option 2: SQLite (For Everything Else)

import sqlite3

db = sqlite3.connect("app.db")
db.execute("""CREATE TABLE IF NOT EXISTS users (
    id INTEGER PRIMARY KEY AUTOINCREMENT,
    name TEXT NOT NULL,
    email TEXT UNIQUE,
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
)""")

# Insert
db.execute("INSERT INTO users (name, email) VALUES (?, ?)", ("Alice", "alice@example.com"))
db.commit()

# Query
for row in db.execute("SELECT * FROM users WHERE name LIKE ?", ("%Ali%",)):
    print(row)

SQLite handles:

Millions of records without breaking a sweat
Concurrent reads (WAL mode)
Full SQL — JOINs, aggregations, indexes
Zero configuration — it's just a file
Built into Python — no pip install needed

Option 3: DuckDB (For Analytics)

import duckdb

# Query CSV files directly — no import needed
result = duckdb.sql("""
    SELECT category, COUNT(*), AVG(price)
    FROM 'products.csv'
    GROUP BY category
    ORDER BY 2 DESC
""")
print(result)

When you need to analyze data but don't need a persistent database. Full DuckDB tutorial →

Option 4: SQLite + Vector Search (For AI Projects)

import sqlite3
import sqlite_vec  # pip install sqlite-vec

db = sqlite3.connect("knowledge.db")
db.enable_load_extension(True)
sqlite_vec.load(db)

# Now you have vector similarity search in SQLite
# No Pinecone, no Weaviate, no Docker

Build RAG apps and semantic search with zero infrastructure. Full tutorial →

When You Actually Need PostgreSQL

Multiple servers writing to the same database
Complex permissions and row-level security
PostGIS for geospatial queries
Full-text search at scale (though SQLite FTS5 is surprisingly good)
Your team already uses it and has the infrastructure

The Side Project Graveyard

I've seen dozens of side projects die at the "set up the database" step:

Choose between PostgreSQL, MySQL, MongoDB
Install Docker (or sign up for a managed service)
Configure connection strings, migrations, ORMs
Realize you need a migration tool
Spend 3 hours on Docker networking
Give up and move to the next idea

With SQLite:

import sqlite3
db = sqlite3.connect("app.db")
Ship it

But What About Deployment?

SQLite works perfectly on:

VPS (DigitalOcean, Hetzner, Fly.io) — the .db file lives next to your app
Serverless — Cloudflare D1, Turso (distributed SQLite)
Desktop apps — Electron, Tauri
Mobile — Every phone already has SQLite

The only place it doesn't work well: horizontal scaling with write-heavy workloads. But if your side project has that problem, congratulations — you have a real business, not a side project.

My Stack for Side Projects

Frontend: HTML + vanilla JS (or React if needed)
Backend: Python (FastAPI) or Node (Hono)
Database: SQLite (always)
Hosting: Fly.io ($0-5/mo) or VPS ($5/mo)
Total monthly cost: $0-5

No Docker. No Kubernetes. No managed database. No $50/month "hobby" tier.

What do you use for side project databases? Am I wrong about PostgreSQL being overkill for most projects? Fight me in the comments.

I write about practical dev tools and keeping things simple. Follow for more.

Need custom dev tools, scrapers, or API integrations? I build automation for dev teams. Email spinov001@gmail.com — or explore awesome-web-scraping.

DEV Community