• A BTP global account If you are an SAP Developer or SAP employee, please refer to the following links ( for internal SAP stakeholders only ) - How to create a BTP Account (internal) SAP AI Core If you are an external developer or a customer or a partner kindly refer to this tutorial
• You have created an SAP AI Core instance.

• How to create a connection between SAP AI Core and your tool of choice (SAP AI Launchpad, Postman, python)
• How to look explore the tools for SAP AI Core

If you want to use SAP AI Core with python, select the tab for SAP AI Core SDK or AI API client SDK.

Please find downloadable sample notebooks for the tutorials : . Note that these tutorials are for demonstration purposes only and should not be used in production environments. To execute them properly, you'll need to set up your own S3 bucket or provision services from BTP, including an AI Core with a standard plan for narrow AI and an extended plan for GenAI HUB. Ensure you input the service keys of these services into the relevant cells of the notebook. Link to notebook

Using your SAP Business Technology Platform (BTP), to generate keys for the SAP AI Core instance, using Provision SAP AI Core

The following is a set of keys for example reference only. Please generate keys for your own SAP AI Core instance.

{
  "clientid": "ss-2222-aaa-4444-888-8555!bbb|aicore!bbb",
  "appname": "2222-aaa-4444-888-8555!bbb|aicore!bbb",
  "identityzone": "tutorial",
  "identityzoneid": "11111-dddd-444-888-888888",
  "clientsecret": "qwertyuiop",
  "serviceurls": {
  "AI_API_URL": "https://api.ai.prod.us-east-1.aws.ml.hana.ondemand.com"
  },
  "url": "https://tutorial.authentication.us10.hana.ondemand.com"
}

[OPTION BEGIN [SAP AI Launchpad]]

Subscribe to SAP AI Launchpad on your SAP BTP Cockpit, and add SAP AI Launchpad to the Entitlements of your subaccount.

Access the app through your SAP BTP Cockpit.

[OPTION END]

[OPTION BEGIN [Postman]]

Download and install Postman.

[OPTION END]

[OPTION BEGIN [SAP AI Core SDK]]

Download python 3.

Start the installer with the following settings.

The command line tool pip is the python package installer, it will be installed automatically with python.

Use pip to install the SAP AI Core SDK.

pip install ai-core-sdk

Use pip to install Jupyter Notebook. Jupyter is used to execute python code pieces in the form of reusable cells.

pip install notebook

[OPTION END]

[OPTION BEGIN [AI API client SDK]]

The command line tool pip is the python package installer, it will be installed automatically with python.

Use pip to install the AI API client SDK.

pip install ai-api-client-sdk

Use pip to install Jupyter Notebook. Jupyter is used to execute python code pieces in the form of reusable cells.

pip install notebook

[OPTION END]

[OPTION BEGIN [SAP AI Launchpad]]

Click on the Workspaces app. Click Add to create connection to your SAP AI Core instance.

Enter the name my-ai-core for the connection . Fill other details from your SAP AI Core service key as marked in the image.

INFORMATION If you find that a connection already exists in your SAP AI Launchpad then you may add your connection under a different name. You may add any number of connections to SAP AI Core instances in your SAP AI Launchpad (this may not be case if you are using Free Tier).

[OPTION END]

[OPTION BEGIN [Postman]]

Visit SAP API Business Hub and search for "SAP AI Core". In this site you can find the latest collections and endpoints available for use with SAP AI Core.

Click on the Download icon next to the JSON. This downloads a file AI_CORE_API.json which has the available API endpoint of SAP AI Core.

On the Postman app, click Import and open the downloaded JSON file.

Click Upload files and open AI_CORE_API.json file. Next, click Show advanced settings. Set the following settings:

Leave other settings unchanged and click Import.

You now have your copy of the official API collections of SAP AI Core named AI Core.

Click on the collection AI Core, this shows you the settings page for the whole collections.

Under the Authorization tab, scroll to Configure New Token. Edit the settings, using the example and instructions below as a guide.

1. Type any descriptive name for Token Name.
2. Set Grant Type to Client Credentials.
3. Paste your url (from your SAP AI Core keys) for Access Token URL. Also notice the suffix /oauth/token, which you need to add.
4. Set value from your keys.
5. Set value from your keys.
6. Click Save.

This process saves your credentials to make the repetitive task of token generation (covered later) easier.

Click on the Variables tab. Fill the settings as mentioned below.

1. Remove the values from the Initial Value column to avoid confusion. Set the value from your AI_API_URL for the variable baseUrl.

Note that you have to set part of your AI_API_URL (from SAP AI Core Keys) for the value region (in Postman), for this you can put any part, there are no specific constraint. But please ensure you add the suffix /v2 at the end of you baseUrl as shown in the screenshot.

2. Click Save.

[OPTION END]

[OPTION BEGIN [SAP AI Core SDK]]

Start your Jupyter notebook. This will start a Jupyter session in the browser.

jupyter notebook

Create a python notebook.

Create a new cell.

Run the cell.

Use this notebook in the tutorial to try out python commands for yourself. Please explore how to save, rename and create new cells in the notebook.

[OPTION END]

[OPTION BEGIN [AI API client SDK]]

Start your Jupyter notebook. This will start a Jupyter session in the browser.

jupyter notebook

Create a python notebook.

Create a new cell.

Run the cell.

Use this notebook in the tutorial to try out python commands for yourself. Please explore how to save, rename and create new cells in the notebook.

[OPTION END]

[OPTION BEGIN [SAP AI Launchpad]]

Click on my-ai-core under the Workspaces app. You will observe the connection name on the header. You can find new apps on the side navigation bar, like SAP AI Core Administration and ML Operations.

Select default from the Resource Groups Pane. Again this selection is updated in the header.

[OPTION END]

[OPTION BEGIN [Postman]]

Click on the collection AI Core. Navigate to the Authorization tab and scroll below. Click Get New Access Token.

IMPORTANT: If you face error in this step, please check previous steps and ensure you added the suffix /oauth/token to the Access Token URL.

Follow the guided steps in the dialog box that appears.

Click Use Token.

[OPTION END]

[OPTION BEGIN [SAP AI Core SDK]]

Edit the command with your SAP AI Core credentials. Please add the suffixes to the values of base_url and auth_url as shown in the code snippet below. Execute the command.

# Load Library
from ai_core_sdk.ai_core_v2_client import AICoreV2Client

# Create Connection
ai_core_client = AICoreV2Client(
    base_url = "<YOUR_AI_API_URL>" + "/v2", # The present SAP AI Core API version is 2
    auth_url=  "<YOUR_url>" + "/oauth/token", # Suffix to add
    client_id = "<YOUR_clientid>",
    client_secret = "<YOUR_clientsecret>"
)

[OPTION END]

[OPTION BEGIN [AI API client SDK]]

Edit the command with your SAP AI Core credentials. Please add the suffixes to the values of base_url and auth_url as shown in the code snippet below. Execute the command.

# Load library
from ai_api_client_sdk.ai_api_v2_client import AIAPIV2Client

# Create connection
ai_api_client = AIAPIV2Client(
    base_url = "https://api.ai.ml.hana.ondemand.com" + "/v2", # The present SAP AI Core API version is 2
    auth_url=  "https://tutorial.authentication.sap.hana.ondemand.com" + "/oauth/token", # Suffix to add
    client_id = "ab-cdefg-111-12233!h7777|aicore!1111",
    client_secret = "qwertyiop="
)

[OPTION END]

[OPTION BEGIN [SAP AI Launchpad]]

Click on Scenarios in the ML Operations app. The term Scenario refers to a use case. You may not have any scenario listed. You will learn how to add workflows (AI pipelines) to SAP AI Core in future tutorials, which can then be used to create scenarios.

[OPTION END]

[OPTION BEGIN [Postman]]

List all the resource groups. Check that there is a resource group named default in the response.

Troubleshooting (Optional)

If you receive the error: RBAC: Access Denied:

1. Click on the collection named AI Core, click the Authorization tab, and check whether a token is being used or not. It might be that you generated a token but forgot to click Use Token in the process.

2. Click the Variables tab and check if the value for baseUrl matches the value of AI_API_URL and you have the suffix /v2 at the end.

[OPTION END]

[OPTION BEGIN [SAP AI Core SDK]]

Execute the command below to list the GitHub repositories connected to your SAP AI Core. If the response prints error stack trace, verify that you used the correct set of keys and added the suffixes mentioned the previous step, when setting the connection.

response = ai_core_client.repositories.query()
print(response.count)

[OPTION END]

[OPTION BEGIN [AI API client SDK]] Execute the command below to list the GitHub repositories connected to your SAP AI Core. If the response prints error stack trace, verify that you used the correct set of keys and added the suffixes mentioned the previous step, when setting the connection.

ai_api_client.rest_client.get(
    path="/admin/repositories"
)

[OPTION END]

[OPTION BEGIN [SAP AI Launchpad]]

If you are don't have apps (like SAP AI Core Administration and ML Operations) on the navigations pane, please check that you are assigned the required role collections in the SAP BTP Cockpit for SAP AI Launchpad.

If you have assigned the roles and the sections are still not visible then please clear your browser cache and relaunch.

INFORMATION: New role collections introduced in new releases of SAP AI Launchpad are not auto-assigned to existing users. The administrator of your SAP BTP Cockpit must ensure that the correct users are assigned new roles. New role collections and other feature releases can be found in Release Notes for SAP AI Launchpad.

[OPTION END]

[OPTION BEGIN [Postman]]

You added the suffix /v2 in your baseUrl, because the current SAP AI Core endpoint version is v2. You will find all the related information in the SAP API Business Hub.

[OPTION END]

[OPTION BEGIN [SAP AI Core SDK]]

Compare the previous step between SAP AI Core SDK and AI API client SDK. Whilst both are python packages offered to control SAP AI Core, the AI API client SDK can also control other AI runtimes which implement the AI API. Therefore the AI API client SDK lacks native function support for operations specific to SAP AI Core, like GitHub repository connection control.

[OPTION END]

[OPTION BEGIN [AI API client SDK]]

When creating a connection to SAP AI Core using the AI API client SDK you must use one of the following:

• Type 1: Required when performing Admin operations, where base_url has suffix /v2

  # Load library
  from ai_api_client_sdk.ai_api_v2_client import AIAPIV2Client
  
  # Create connection
  ai_api_client = AIAPIV2Client(
      base_url = "https://api.ai.ml.hana.ondemand.com" + "/v2", # The present SAP AI Core API version is 2
      auth_url=  "https://tutorial.authentication.sap.hana.ondemand.com" + "/oauth/token", # Suffix to add
      client_id = "ab-cdefg-111-12233!h7777|aicore!1111",
      client_secret = "qwertyiop="
  )

• Type 2: Required when performing lifecycle management operations, where base_url has suffix /v2/lm

  # Load library
  from ai_api_client_sdk.ai_api_v2_client import AIAPIV2Client
  
  # Create connection
  ai_api_client = AIAPIV2Client(
      base_url = "https://api.ai.ml.hana.ondemand.com" + "/v2/lm", # The present SAP AI Core API version is 2
      auth_url=  "https://tutorial.authentication.sap.hana.ondemand.com" + "/oauth/token", # Suffix to add
      client_id = "ab-cdefg-111-12233!h7777|aicore!1111",
      client_secret = "qwertyiop="
  )

Visit SAP API Business Hub to find out which operation endpoints start with /v2/admin (Admin operations) and which ones start with /v2/lm (lifecycle management).

[OPTION END]

[OPTION BEGIN [Postman]]

If Postman is unable to decipher the schema of the JSON file that you downloaded, you will require the following information, which is stored, for your reference, at API Business Hub of SAP AI Core.

1. Click on any of the endpoints you would like to know information on, for example, click on Artifact.

2. Click on POST endpoint.

3. Click the Schema tab. Here, you will find the detailed description of the values supported by this endpoint.

[OPTION END]

[OPTION BEGIN [SAP AI Core SDK]]

1. Use python's dir function to list contained modules and functions.

print(dir(ai_core_client))

2. Use python's help function on a listed module or function to get a docstring, which documents what the function does and what parameters it requires.

help(ai_core_client.scenario.query)

[OPTION END]

[OPTION BEGIN [AI API client SDK]]

1. Use python's dir function to list contained modules and functions.

print(dir(ai_api_client))

2. Use python's help function on a listed module or function to get a docstring, which documents what the function does and what parameters it requires.

help(ai_api_client.scenario.query)

[OPTION END]

For support, look at the Troubleshooting Guide for SAP AI Core. If this does not help resolve your issue, please raise a ticket.