Snowflake Cortex Agents Python SDK

Python SDK for Snowflake Cortex Agents and Cortex Analyst, with sync and async clients, SSE streaming, and chart helpers for agent responses.

Installation

Install from PyPI:

pip install snowflake-cortex-agents

Optional extras:

• dotenv for .env loading
• charts for Altair and Pandas chart rendering
• all for all optional dependencies

pip install snowflake-cortex-agents[all]

For source and contributor setup, see CONTRIBUTING.md.

Credentials

Set credentials with environment variables:

export SNOWFLAKE_ACCOUNT_URL=https://your-account.snowflakecomputing.com
export SNOWFLAKE_PAT=your-personal-access-token

Or install snowflake-cortex-agents[dotenv] and use a .env file:

SNOWFLAKE_ACCOUNT_URL=https://your-account.snowflakecomputing.com
SNOWFLAKE_PAT=your-personal-access-token

When working from this repository, you can copy .env.example to .env and fill in your values.

Quick Start

Both Cortex Agent run() and Cortex Analyst message() return streaming SSE responses.

Run an existing agent

from cortex_agents import CortexAgent

with CortexAgent() as client:
    response = client.run(
        "What's this month revenue?",
        agent_name="MY_AGENT",
        database="MY_DATABASE",
        schema="MY_SCHEMA",
    )

    for event in response:
        if event["type"] == "text.delta":
            print(event["data"]["text"], end="", flush=True)

Ask Cortex Analyst a question

from cortex_agents import CortexAnalyst

with CortexAnalyst() as analyst:
    response = analyst.message(
        question="What were the top 5 products by revenue last month?",
        semantic_model_file="@my_stage/semantic_model.yaml",
    )

    for event in response:
        if event["type"] == "text.delta":
            print(event["data"]["text"], end="", flush=True)
        elif event["type"] == "sql.delta":
            print(event["data"]["sql"], end="", flush=True)

Async support

Async clients are available via AsyncCortexAgent and AsyncCortexAnalyst. See docs/quickstart.md and examples/example_agent_async.py for runnable async examples.

Documentation

• docs/installation.md: installation, optional extras, and credential setup
• docs/quickstart.md: sync and async getting-started flows
• docs/guides/agents.md: Cortex Agent usage
• docs/guides/analyst.md: Cortex Analyst usage
• docs/guides/threads_api.md: thread management
• docs/guides/agents_threads.md: conversational agent patterns
• USAGE.md: broader API usage notes
• docs/guides/chart_plotting.md: chart rendering helpers
• examples/: runnable examples

Requirements

• Python 3.10+
• Snowflake account with Cortex enabled
• Personal Access Token (PAT)

Contributing

Development workflow, uv-based setup, and quality checks are documented in CONTRIBUTING.md.

License

This project is licensed under the MIT License. See LICENSE.