Installing Sema4.ai Studio and SDK

Before building your first agent, you'll need to install and configure a few tools. This guide walks you through installing Sema4.ai Studio and the Sema4.ai SDK.

Sema4.ai Studio: This is the primary environment for building AI agents. It provides a user-friendly interface for business process owners and developers to create, test, and deploy agents using natural language and code.
Sema4.ai SDK: The Software Development Kit (SDK) is a set of development tools and libraries that allow you to extend the capabilities of your AI agents. It enables you to create custom actions, integrate with various data sources, and build more complex functionalities for your agents.

Before you start

Make sure you have the following:

Your Snowflake account has support for required features:
- Needs to be an AWS region (other cloud providers will be supported later)
- Region has support (opens in a new tab) for claude-3-5-sonnet model, or the account has cross-region inference enabled (opens in a new tab)
- Snowpark Container Services are available (opens in a new tab) - please note they are not enabled for trial accounts
A Snowflake account with access to the Sema4.ai Team Edition app
A Snowflake user with the ACCOUNTADMIN role
Snowflake key-pair authentication is configured for your Snowflake user, and you have a private key file.

Current (1.3.xx) version only supports Snowflake key-pair authentication. Follow steps here (opens in a new tab) to set up private key-pair authentication for your Snowflake account. Additional authentication methods will be supported soon.

Windows users only: when setting up Snowflake key-pair authentication, please notice that OpenSSL is not included by default on Windows operating systems and requires separate installation.

Windows users who have already installed for example Git can utilize the Git Bash application, which includes OpenSSL functionality, to run the commands specified in the above guide. Alternatively, OpenSSL can be downloaded and installed directly from the official website or trusted third-party providers. Another option is to use Windows Subsystem for Linux (WSL), which provides a Linux environment where OpenSSL is readily available or easily installable.

Install Sema4.ai Studio

Download and install Sema4.ai Studio from the link that you can find in the Control Room or that has been provided in your invitation email or by your Sema4.ai point of contact. The application is available for both macOS and Windows.

You can also watch our tutorial video here (opens in a new tab) to see how to link your Snowflake account to Sema4.ai Studio, including how to set up key-pair authentication.

Install

Double-click the installer and follow the on-screen instructions
Accept the End User License Agreement (EULA)
Launch Sema4.ai Studio from your Applications folder (macOS) or Start menu (Windows)
Wait for the initial environment setup to complete (first launch will take several minutes)
You will see a welcome screen like this:

Link to Snowflake

By linking Studio to your Snowflake account you'll get access to resources such as Cortex AI LLM models, your Snowflake databases and services. Click Link Snowflake and follow the steps below.

Enter your Snowflake account details:
- Account identifier (e.g., ZVZWMYO-HP00956)
- Username
- Role
Click Select your private key to open a file picker to find and select the private key file you created when setting up Snowflake key-pair authentication.

On macOS pressing Command+Shift+. will change to the mode which reveals all folders, including the ones named for example .ssh where the private key file is typically located.

If your private key file has a passphrase, enter it to the last field.
Click Link to Snowflake button.

Your connection gets verified, and if all works well, you'll see a message confirming that you're connected. Your Studio is now ready to use!

Sema4.ai Studio - Ready to build agents!

Install Sema4.ai SDK

If you are planning to build custom actions or named queries, you'll need to install Sema4.ai SDK. Sema4.ai SDK is available as a VS Code extension, and it also works with Cursor. The screenshots in this tutorial are from Cursor, but the steps are the same for VS Code.

Install IDE and extensions

Download and install VS Code (opens in a new tab) or Cursor (opens in a new tab) IDE unless you already have one installed
Install two Sema4.ai extensions either from the marketplace links below, or by searching for Sema4.ai within the extensions tab in the IDE.

Search for Sema4.ai in the extension view

Links to the extensions:

Sema4.ai SDK (opens in a new tab) (this is the main extension for building agents and actions)
Sema4.ai Data Access (opens in a new tab) (this manages connectivity to data sources)

Verify installation

Once installed, make sure you open your IDE in an empty directory, then navigate to the SDK extension in the navigation bar. You should see a view like this:

Admin guide Agent templates