CreateAISession()

edit

Creates a new AI session for interaction with Large Language Models (LLMs).

The session maintains conversation history and context between interactions.

Each session can be configured with different parameters to optimize for specific use cases like code analysis, content generation, or data processing.

Common usage patterns:

  • Exception analysis and debugging assistance
  • Code documentation generation
  • Query optimization suggestions
  • Security review assistance
  • Performance optimization recommendations

The session persists until explicitly terminated or the application restarts.

CreateAISession( name=string, systemMessage=string, limit=numeric, temperature=numeric );

Returns: object

Argument Description Default
name
string, required 		edit

Specifies which AI endpoint configuration to use. Can be provided in two formats:

  1. Direct endpoint name:

The name of an endpoint as defined in the Lucee Administrator (similar to how datasource names work)

  1. Default reference:

Using the format "default:category" to use the endpoint configured as the default for that specific category in the Lucee Administrator.

Currently supported default categories:

  • exception: For exception analysis
  • documentation: For documentation tasks

The endpoint configurations and their default category assignments are managed in the Lucee Administrator.

Alias: aiName, nameAI
systemMessage
string, optional 		edit

Initial instruction set that defines the AI's behavior and expertise for this session. This message sets the context and rules for all subsequent interactions.

Best practices:

  • Be specific about the AI's role and expertise
  • Define output format requirements clearly
  • Include any necessary constraints or rules
  • Specify error handling preferences
  • Define response structure expectations

The system message persists throughout the session and influences all responses.

Alias: initalMesssage, message
limit
numeric, optional 		edit

Maximum number of question-answer pairs to keep in the conversation history. Once reached, older messages will be removed.

Consider:

  • Higher limits provide more context but consume more memory
  • Lower limits are more memory efficient but may lose important context
  • For complex analysis tasks, consider limits of 10-20
  • For simple Q&A, limits of 5-10 may suffice

Alias: conversationHistoryLimit

50
temperature
numeric, optional 		edit

Controls response randomness (0.0 to 1.0). Lower values make responses more focused and deterministic, higher values make them more creative and varied.

Recommended settings:

  • 0.0-0.3: Technical analysis, debugging, code review
  • 0.3-0.5: Documentation generation, error explanations
  • 0.5-0.7: General purpose interactions
  • 0.7-1.0: Creative content generation

For exception analysis and debugging, lower temperatures (0.2-0.3) are recommended for more consistent and precise responses.

0.7

Examples

edit 
// Create an AI session with the specified service name, e.g., "test_ai"
aisession = CreateAISession("test_ai");

See also