opendbpylot (Python)

Python bindings for opendbpylot — turn a natural-language question into SQL, run it on your database, and get the results. Uses Retrieval-Augmented Generation (RAG) with a self-repairing SQL loop.

pip install opendbpylot

This installs both a Python library and a fully-working dbpylot command — the whole CLI runs in-process from the native module, so no Rust toolchain is needed:

dbpylot init        # setup wizard
dbpylot             # chat with your database
dbpylot serve       # web UI
dbpylot ask "how many orders per country?"
dbpylot mcp         # serve as an MCP server for agent hosts (OpenPylot, Claude Desktop, …)

dbpylot mcp exposes the engine over the Model Context Protocol (stdio). Secrets can be set non-interactively for scripted setups: printf '%s' "$KEY" | dbpylot config set-key openai and dbpylot config set-db sqlite /data/app.db.

Usage (library)

First configure an LLM provider + a database. The bindings share the same config as the dbpylot CLI (install it with cargo install opendbpylot), so run the wizard once:

dbpylot init
# …or, from Python:
python -c "import opendbpylot; opendbpylot.OpenDbPylot.init()"

Then:

import opendbpylot

bot = opendbpylot.OpenDbPylot()
result = bot.ask("how many orders per country?")

print(result["sql"])          # the generated (and self-repaired) SQL
print(result["columns"])      # column names
for row in result["rows"]:    # result rows
    print(row)

# Teach it about your schema / business rules:
bot.train_documentation("Revenue excludes cancelled and refunded orders.")
bot.train_question_sql("top products", "SELECT name FROM products ORDER BY price DESC LIMIT 10;")

Fully self-contained web UI

You don't even need the CLI — launch the embedded web app straight from Python and configure everything (LLM + database) in the browser Settings panel:

import opendbpylot
opendbpylot.OpenDbPylot.serve()   # opens http://127.0.0.1:8080, runs until interrupted

API

• OpenDbPylot() — load the engine from your saved configuration.
• .ask(question) -> dict — {sql, columns, rows, repairs_used, answer?}.
• .train_ddl(ddl), .train_documentation(doc), .train_question_sql(q, sql).
• OpenDbPylot.serve() — run the embedded web UI in-process (no CLI binary needed).
• OpenDbPylot.init() / .doctor() — convenience shims that call the dbpylot CLI (install it with cargo install opendbpylot); serve() + Settings does the same setup.

Licensed under Apache-2.0.