pgvector-zerodb

Drop-in pgvector wrapper with auto-provisioned Postgres. Zero config vector database.

pip install pgvector-zerodb

Quick Start

from pgvector_zerodb import auto_connect

# Automatically provisions a free Postgres with pgvector if no DATABASE_URL is set
conn = auto_connect()

cur = conn.cursor()
cur.execute("CREATE TABLE items (id serial PRIMARY KEY, embedding vector(3))")
cur.execute("INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]')")
conn.commit()

cur.execute("SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5")
print(cur.fetchall())

That's it. No database setup, no config files, no cloud console.

How It Works

1. You call auto_connect()
2. It checks for DATABASE_URL or PGVECTOR_ZERODB_URL environment variable
3. If neither exists, it provisions a free managed Postgres with pgvector via AINative
4. Credentials are saved to ~/.ainative/pgvector-config.json for reuse
5. Returns a standard psycopg2 connection with pgvector types registered

Use Your Own Postgres

If you already have a Postgres instance with pgvector:

export DATABASE_URL="postgresql://user:pass@host:5432/dbname"

from pgvector_zerodb import auto_connect
conn = auto_connect()  # Uses your DATABASE_URL, no provisioning

Same pgvector API

Everything from pgvector is re-exported. Existing code works unchanged:

from pgvector_zerodb import register_vector
import psycopg2

conn = psycopg2.connect("postgresql://...")
register_vector(conn)

With SQLAlchemy

from pgvector_zerodb import get_connection_string
from sqlalchemy import create_engine

engine = create_engine(get_connection_string())

With Django

Add to settings.py:

from pgvector_zerodb import get_connection_string
import dj_database_url

DATABASES = {"default": dj_database_url.parse(get_connection_string())}

Environment Variables

Free Tier

The auto-provisioned Postgres includes:

• pgvector extension pre-installed
• 256 MB storage
• Shared compute
• No credit card required

Upgrade at ainative.studio for dedicated resources.

License

MIT