Skip to main content
Version: v5.2

Support for different LLMs

Overview#

Platform version 5.2 introduced the ability to define an agent using your choice of LLM provider.

The current two options are:

  • OpenAI (GPT model)
  • Anthropic (Claude model)

Defining an agent with a specific LLM#

You can choose which AI model to use when you are defining a new agent. These choices are set using the provider and model fields. Refer to the summary of these fields below.

FieldDescription
providerDetermines the LLM provider, currently accepts openai or anthropic.
modelDetermines the specific AI model to use from a particular provider. For example, for anthropic, it could be claude-opus-4-6.
temperatureA numeric setting which determines the randomness and predictability of an AI model's text responses.

How to define an agent which uses OpenAI#

Refer to the example below which defines a new agent which uses OpenAI as the LLM provider and specifies that the agent uses OpenAI's GPT 4.0 version model. 


 {        "_name": "Related Query Helper Agent",        "_background": "It interacts with Related Query Tool to generate related queries as JSON based on user prompt and provide this as input to the SearchRelatedItemsTool to retrive data from Item service",        "_userType": "related_query_helper_agent",        "_namespaces": [            "iputGraph_kxYOcW7O"        ],        "_config": {            "_model": "gpt-4o",            "_provider": "openai",            "_temperature": 1.5        },        "_tools": [            "RelatedQueryTool",            "SearchRelatedItemsTool"        ] }

How to define an agent which uses Anthropic#

Refer to the example below which defines a new agent which uses Anthropic as the LLM provider and specifies that the agent uses Anthropic's Claude Opus 4.6 version model. 

 {   "_name": "File Specialist",   "_userType": "file_specialist",   "_description": "Answers file-service questions using the file tools: list/get files, folder trees, versions, preview/download URLs, upload metadata and file permissions.",   "_background": "You are the File Specialist for a digital-twin platform. Answer ONLY the file-service part of the user's request, using your tools to fetch real data: listing and retrieving files, walking folder trees, file versions, preview and download URLs, upload metadata and file permissions.",   "_config": {    "_provider": "anthropic",    "_model": "claude-opus-4-6",    "_temperature": 1   },   "_tools": [    "GetFilesTool",    "GetFileTool",    "GetAllFilesInTreeTool",    "GetFileUrlTool",    "GetFilePreviewUrlTool",    "GetFileVersionsTool",    "GetFileVersionTool",    "GetFileInfoWithUploadMetaTool",    "GetFilePermissionsTool"   ],   "_knowledgebases": [],   "_invocationMessage": "Looking up files...",   "_namespaces": ["ns"]}

REST API support#

The AI Service REST API has two API calls specifically to support multiple LLM providers.

  • LLM Models

  • Providers

LLM Models#

The LLM Models API call lists supported LLM providers and their available model names. For more information, see the LLM Models page.

Providers#

The Providers API call lists the supported LLM provider keys configured for the AI Service. For more information, see the Providers page.