SAP Cloud SDK for AI is the official Software Development Kit (SDK) for SAP AI Core, SAP Generative AI Hub, and Orchestration Service.
This package incorporates generative AI foundation models into your AI activities in SAP AI Core and SAP AI Launchpad.
- Table of Contents
- Installation
- Prerequisites
- Relationship between Models and Deployment ID
- Usage
- Local Testing
- Support, Feedback, Contribution
- License
$ npm install @sap-ai-sdk/foundation-models
- Enable the AI Core service in SAP BTP.
- Configure the project with Node.js v20 or higher and native ESM support.
- Ensure a deployed OpenAI model is available in the SAP Generative AI Hub.
- Use the
DeploymentApi
from@sap-ai-sdk/ai-api
to deploy a model. Alternatively, you can also create deployments using the SAP AI Launchpad. Deployment can be set up for each model and model version, as well as a resource group. - Once a deployment is complete, access the model via the
deploymentUrl
.
- Use the
Accessing the AI Core Service via the SDK
The SDK automatically retrieves the
AI Core
service credentials and resolves the access token needed for authentication.
- In Cloud Foundry, it's accessed from the
VCAP_SERVICES
environment variable.- In Kubernetes / Kyma environments, you have to mount the service binding as a secret instead, for more information refer to this documentation.
SAP AI Core manages access to generative AI models through the global AI scenario foundation-models
.
Creating a deployment for a model requires access to this scenario.
Each model, model version, and resource group allows for a one-time deployment.
After deployment completion, the response includes a deploymentUrl
and an id
, which is the deployment ID.
For more information, see here.
Resource groups represent a virtual collection of related resources within the scope of one SAP AI Core tenant.
Consequently, each deployment ID and resource group uniquely map to a combination of model and model version within the foundation-models
scenario.
You can pass the model name as a parameter to a client, the SDK will implicitly fetch the deployment ID for the model from the AI Core service and use it in the request.
By default, the SDK caches the deployment information, including the deployment ID, model name, and version, for 5 minutes to avoid performance issues from fetching this data with each request.
import {
AzureOpenAiChatClient,
AzureOpenAiEmbeddingClient
} from '@sap-ai-sdk/foundation-models';
// For a chat client
const chatClient = new AzureOpenAiChatClient({ modelName: 'gpt-4o' });
// For an embedding client
const embeddingClient = new AzureOpenAiEmbeddingClient({ modelName: 'gpt-4o' });
The deployment ID and resource group can be used as an alternative to the model name for obtaining a model.
const chatClient = new AzureOpenAiChatClient({
deploymentId: 'd1234',
resourceGroup: 'rg1234'
});
Use the AzureOpenAiChatClient
to send chat completion requests to an OpenAI model deployed in SAP generative AI hub.
The client sends request with Azure OpenAI API version 2024-06-01
.
import { AzureOpenAiChatClient } from '@sap-ai-sdk/foundation-models';
const chatClient = new AzureOpenAiChatClient('gpt-4o');
const response = await chatClient.run({
messages: [
{
role: 'user',
content: 'Where is the deepest place on earth located'
}
]
});
const responseContent = response.getContent();
Multiple messages can be sent in a single request, enabling the model to reference the conversation history.
Include parameters like max_tokens
and temperature
in the request to control the completion behavior:
const response = await chatClient.run({
messages: [
{
role: 'system',
content: 'You are a friendly chatbot.'
},
{
role: 'user',
content: 'Hi, my name is Isa'
},
{
role: 'assistant',
content:
'Hi Isa! It is nice to meet you. Is there anything I can help you with today?'
},
{
role: 'user',
content: 'Can you remind me, What is my name?'
}
],
max_tokens: 100,
temperature: 0.0
});
const responseContent = response.getContent();
const tokenUsage = response.getTokenUsage();
logger.info(
`Total tokens consumed by the request: ${tokenUsage.total_tokens}\n` +
`Input prompt tokens consumed: ${tokenUsage.prompt_tokens}\n` +
`Output text completion tokens consumed: ${tokenUsage.completion_tokens}\n`
);
Refer to AzureOpenAiChatCompletionParameters
interface for other parameters that can be passed to the chat completion request.
Use the AzureOpenAiEmbeddingClient
to send embedding requests to an OpenAI model deployed in SAP generative AI hub.
import { AzureOpenAiEmbeddingClient } from '@sap-ai-sdk/foundation-models';
const embeddingClient = new AzureOpenAiEmbeddingClient(
'text-embedding-ada-002'
);
const response = await embeddingClient.run({
input: 'AI is fascinating'
});
const embedding = response.getEmbedding();
Set custom request configuration in the requestConfig
parameter when calling the run()
method of a chat or embedding client.
const response = await client.run(
{
...
},
{
headers: {
'x-custom-header': 'custom-value'
// Add more headers here
},
params: {
// Add more parameters here
}
// Add more request configuration here
}
);
For local testing instructions, refer to this section.
This project is open to feature requests, bug reports and questions via GitHub issues.
Contribution and feedback are encouraged and always welcome. For more information about how to contribute, the project structure, as well as additional contribution information, see our Contribution Guidelines.
The SAP Cloud SDK for AI is released under the Apache License Version 2.0.