AnnaAgent Generator

The AnnaAgent Generator creates character files specifically designed for the AnnaAgent client method, which supports multi-session counseling with dynamic memory evolution.

Overview

Property	Value
Key	annaAgent
Type	LLM-based
Output	AnnaAgent character files

Description

The AnnaAgent Generator produces detailed character specifications that are compatible with the AnnaAgent client method. It creates profiles with rich backstories, presenting problems, and memory structures that can evolve across multiple therapy sessions.

Key Features

• Multi-session Support - Generates characters designed for longitudinal therapy
• Memory Evolution - Creates initial memory states that can dynamically update
• Rich Backstories - Detailed personal histories and context
• Session Goals - Defines therapeutic objectives for each character

Configuration

YAML Configuration

generator:
  type: annaAgent
  config:
    model: gpt-4o
    temperature: 0.8
    include_memory_structure: true

Python Usage

from patienthub.generators import GeneratorRegistry

generator = GeneratorRegistry.create("annaAgent", config={
    "model": "gpt-4o",
    "temperature": 0.8
})

character = generator.generate({
    "disorder": "depression",
    "severity": "moderate",
    "session_count": 5
})

Parameters

Parameter	Type	Default	Description
model	string	gpt-4o	The LLM model to use
temperature	float	0.8	Controls creativity in generation
include_memory_structure	bool	true	Include dynamic memory initialization

Output Format

{
  "name": "Sarah",
  "age": 28,
  "background": "Software engineer dealing with work-related stress...",
  "presenting_problem": "Difficulty sleeping, persistent sadness...",
  "memory": {
    "core_beliefs": ["I'm not good enough"],
    "recent_events": [...],
    "therapeutic_progress": []
  },
  "session_goals": [...]
}

Use Cases

• Creating diverse client pools for multi-session research
• Generating training data for AnnaAgent clients
• Building character libraries for educational purposes