AI Auto Value Mapping (Preview Feature)
An optional AI mapping tool (enabled in global settings) streamlines value mapping for enum fields using LLMs (Large Language Models).
Connect integrates with all LLM providers that support an OpenAI-compatible REST API for completion. This includes local and lab Llama via the Llamafile project, and proxies such as openai-api-2-gemini-api.
The AI-enhanced mapping provides a more thorough mapping of enum fields by applying knowledge attained through integrated LLMs.
Note: Preview features may have limitations and differ from finalized versions. They may be discontinued in future releases and are disabled by default, requiring manual activation.
Configuration
To enable AI value mapping in your deployment, contact support or instruct your admin to run the mfcFullRestClient script to set the following properties:
| Property | Description |
|---|---|
llm.integration.enabled
|
Enables or disables AI-powered value mapping. Set to true or false. |
llm.integration.server.chat.url
|
API endpoint URL for the LLM provider (OpenAI, Llamafile server, or proxy). |
llm.integration.server.token
|
Optional authentication token for accessing the LLM provider's API. |
llm.integration.model
|
The LLM model name. |
Commands to enable AI auto value mapping
Use the following commands to enable the AI mapping:
Perform AI auto value mapping
Once setup is complete, you can run the auto mapping.
To perform the mapping:
- In the connection wizard, go to Types and Fields tab.
- Expand a type mapping to view fields.
- Click Edit Value Map to open the Value Maps dialog.
- Click Auto Value Mapping by AI to generate automatic mapping of field values.
See also:

