Database Queries¶

SQL from your AI. Three functions. Any database.

The db.* pack gives LLMs direct database access - explore tables, inspect schemas, run queries.

Available Functions¶

Function	Purpose
`db.tables(project)`	List all tables in a database
`db.schema(project, table)`	Get table schema
`db.query(project, sql)`	Execute SQL query

Setup¶

Configure database connections in ot-serve.yaml:

projects:
  myapp:
    path: /path/to/project
    attrs:
      db_url: postgresql://localhost/myapp

  demo:
    path: .
    attrs:
      db_url: sqlite:///demo/db/northwind.db

Basic Usage¶

List Tables¶

db.tables(project="demo")

Get Schema¶

db.schema(project="demo", table="customers")

Run Query¶

db.query(project="demo", sql="SELECT * FROM customers LIMIT 10")

Configuration¶

Setting	Default	Description
`db.max_chars`	4000	Maximum output characters

Override in config:

tools:
  db:
    max_chars: 8000  # Larger output

Supported Databases¶

Any SQLAlchemy-compatible database:

SQLite: sqlite:///path/to/db.db
PostgreSQL: postgresql://user:pass@host/db
MySQL: mysql://user:pass@host/db

Security Notes¶

Queries are read-only by default
Use parameterized queries for user input
Configure max_chars to prevent excessive output

Example Workflow¶

# 1. Explore available tables
tables = db.tables(project="demo")

# 2. Check schema
schema = db.schema(project="demo", table="orders")

# 3. Query data
results = db.query(
    project="demo",
    sql="SELECT customer_id, COUNT(*) as order_count FROM orders GROUP BY customer_id"
)