komodo 3.4.1

Marmot Development Kit — Python SDK and CLI for the Komodo Health platform (Snowflake access, OAuth, MCP, and more).

Marmot Development Kit

The komodo PyPI package is the Python SDK and CLI component of the Marmot Development Kit for the Komodo Health platform. It ships three main surfaces:

Surface	What you get
SDK	Python APIs for OAuth-backed sessions, account context, and Snowflake access via a DB-API 2.0 connection
CLI	The komodo command-line tool for login, account selection, ad hoc SQL, running the MCP server, and more
MCP tools	A Model Context Protocol server you can attach to compatible AI clients so assistants can call Komodo-backed tools (see the docs for configuration)

Together they cover authentication, querying your Komodo data warehouse, and optional assistant-driven workflows.

Features

• OAuth 2.0 device flow — browser-based login and persisted credentials
• Snowflake access — query your warehouse through the platform proxy (SDK or CLI)
• Sync and async queries — blocking execution and async submit/poll helpers in Python
• Automatic token refresh
• Environments — integration or production via KOMODO_ENVIRONMENT

Prerequisites

• Python 3.11 or higher (see requires-python in this package)
• uv (recommended)

Installation

From a new directory, initialize a uv project and add the package:

uv init
uv add komodo

Confirm the CLI:

uv run komodo --version

Quick start

1. Log in (opens a browser):

   uv run komodo login
   

2. Choose your Komodo account:

   uv run komodo account set
   

3. Use the SDK from Python — create first_query.py:

   from komodo import get_snowflake_connection
   
   conn = get_snowflake_connection()
   cursor = conn.cursor()
   cursor.execute("SELECT 1")
   print(cursor.fetchone())
   conn.close()
   

   uv run first_query.py
   

Documentation

Full guides for the Marmot Development Kit—SDK, CLI, MCP setup, authentication (including machine-to-machine), pandas, and API reference:

https://docs.komodohealth.com