"""
Pattern 075: Creative Agent
Description:
A Creative Agent is a specialized AI agent designed to generate original creative
content across various domains including writing (stories, poetry, scripts), visual
concepts (art descriptions, design ideas), music concepts, and other creative works.
This pattern demonstrates how to build an intelligent agent that can understand
creative requirements, generate diverse creative outputs, adapt styles, incorporate
feedback, and maintain creative consistency.
The agent combines creative ideation with structured generation techniques to produce
high-quality, original content. It can work across multiple creative domains, adapt
to different styles and tones, generate variations, and iteratively refine outputs
based on feedback.
Components:
1. Ideation Engine: Generates creative concepts and ideas
2. Style Adapter: Adapts content to different styles and tones
3. Content Generator: Creates structured creative outputs
4. Variation Creator: Produces diverse alternatives
5. Constraint Respector: Adheres to creative constraints
6. Feedback Integrator: Incorporates critique and improvements
7. Consistency Maintainer: Ensures coherent creative vision
8. Multi-Modal Orchestrator: Coordinates different creative types
Key Features:
- Cross-domain creative generation (text, visual concepts, music ideas)
- Style and tone adaptation (formal, casual, poetic, technical)
- Character and world-building for narratives
- Creative constraint handling (theme, length, format)
- Iterative refinement with feedback
- Variation generation for exploration
- Originality assessment
- Creative inspiration from prompts
- Collaborative creativity support
- Genre and format flexibility
Use Cases:
- Story and novel writing
- Poetry and lyric generation
- Screenplay and script writing
- Marketing copy and slogans
- Product naming and branding
- Art concept development
- Game narrative design
- Content marketing
- Creative brainstorming
- Educational creative writing
LangChain Implementation:
Uses ChatOpenAI with varying temperatures for different creative tasks,
specialized chains for ideation, generation, and refinement.
"""
import os
import re
from typing import List, Dict, Any, Optional
from dataclasses import dataclass, field
from enum import Enum
from datetime import datetime
from dotenv import load_dotenv
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.output_parsers import StrOutputParser
from langchain_openai import ChatOpenAI
load_dotenv()
class CreativeType(Enum):
"""Types of creative content"""
STORY = "story"
POEM = "poem"
SCRIPT = "script"
SONG_LYRICS = "song_lyrics"
ARTICLE = "article"
MARKETING_COPY = "marketing_copy"
PRODUCT_NAME = "product_name"
SLOGAN = "slogan"
DIALOGUE = "dialogue"
WORLD_BUILDING = "world_building"
class CreativeStyle(Enum):
"""Creative styles and tones"""
FORMAL = "formal"
CASUAL = "casual"
POETIC = "poetic"
HUMOROUS = "humorous"
DRAMATIC = "dramatic"
MINIMALIST = "minimalist"
ELABORATE = "elaborate"
MYSTERIOUS = "mysterious"
ROMANTIC = "romantic"
DARK = "dark"
class Genre(Enum):
"""Creative genres"""
FANTASY = "fantasy"
SCIENCE_FICTION = "science_fiction"
MYSTERY = "mystery"
ROMANCE = "romance"
HORROR = "horror"
THRILLER = "thriller"
COMEDY = "comedy"
DRAMA = "drama"
ADVENTURE = "adventure"
HISTORICAL = "historical"
@dataclass
class CreativePrompt:
"""Input prompt for creative generation"""
type: CreativeType
theme: str
style: CreativeStyle
genre: Optional[Genre] = None
constraints: Dict[str, Any] = field(default_factory=dict)
inspiration: List[str] = field(default_factory=list)
target_audience: Optional[str] = None
@dataclass
class Character:
"""Character for narrative works"""
name: str
role: str # protagonist, antagonist, supporting
personality: List[str]
background: str
goals: List[str]
conflicts: List[str]
@dataclass
class Setting:
"""Setting for creative works"""
location: str
time_period: str
atmosphere: str
key_features: List[str]
description: str
@dataclass
class CreativeWork:
"""Generated creative content"""
work_id: str
type: CreativeType
title: str
content: str
style: CreativeStyle
genre: Optional[Genre]
characters: List[Character] = field(default_factory=list)
setting: Optional[Setting] = None
themes: List[str] = field(default_factory=list)
word_count: int = 0
metadata: Dict[str, Any] = field(default_factory=dict)
created_at: datetime = field(default_factory=datetime.now)
@dataclass
class CreativeFeedback:
"""Feedback on creative work"""
work_id: str
strengths: List[str]
weaknesses: List[str]
suggestions: List[str]
overall_score: float # 0.0-1.0
specific_notes: Dict[str, str] = field(default_factory=dict)
@dataclass
class CreativeVariation:
"""Variation of a creative work"""
original_id: str
variation_id: str
content: str
changes_made: List[str]
variation_type: str # tone, structure, perspective, etc.
class CreativeAgent:
"""
Agent for generating creative content across multiple domains.
This agent can generate stories, poems, scripts, marketing copy, and other
creative works while adapting to styles, genres, and constraints.
"""
def __init__(self):
"""Initialize the creative agent with specialized LLMs"""
# Ideation engine for brainstorming
self.ideation_llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0.9)
# Generator for main content creation
self.generator_llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0.8)
# Refiner for improving content
self.refiner_llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0.6)
# Critic for evaluation
self.critic_llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0.3)
# Creative works repository
self.works: Dict[str, CreativeWork] = {}
def generate_ideas(
self,
theme: str,
creative_type: CreativeType,
count: int = 5
) -> List[str]:
"""
Generate creative ideas based on theme.
Args:
theme: Central theme or concept
creative_type: Type of creative work
count: Number of ideas to generate
Returns:
List of creative ideas
"""
prompt = ChatPromptTemplate.from_messages([
("system", """You are a creative ideation expert. Generate original and
imaginative ideas based on the theme and content type.
Provide {count} ideas, one per line starting with a number."""),
("user", """Theme: {theme}
Content Type: {type}
Generate {count} creative ideas.""")
])
chain = prompt | self.ideation_llm | StrOutputParser()
try:
response = chain.invoke({
"theme": theme,
"type": creative_type.value,
"count": count
})
ideas = []
for line in response.split('\n'):
line = line.strip()
if line and (line[0].isdigit() or line.startswith('-')):
# Remove numbering
idea = re.sub(r'^\d+[\.\)]\s*', '', line)
idea = re.sub(r'^[-*]\s*', '', idea)
if idea:
ideas.append(idea)
return ideas[:count] if ideas else [f"Idea about {theme}" for _ in range(count)]
except Exception as e:
return [f"Creative idea {i+1} about {theme}" for i in range(count)]
def create_characters(
self,
story_concept: str,
character_count: int = 3
) -> List[Character]:
"""
Create characters for a narrative work.
Args:
story_concept: Brief story concept
character_count: Number of characters to create
Returns:
List of Character objects
"""
prompt = ChatPromptTemplate.from_messages([
("system", """You are a character development expert. Create compelling
characters for the story concept.
For each character, provide:
NAME: character name
ROLE: protagonist/antagonist/supporting
PERSONALITY: traits (comma-separated)
BACKGROUND: brief background
GOALS: goals (comma-separated)
CONFLICTS: conflicts (comma-separated)
---"""),
("user", """Story Concept: {concept}
Create {count} characters.""")
])
chain = prompt | self.ideation_llm | StrOutputParser()
try:
response = chain.invoke({
"concept": story_concept,
"count": character_count
})
characters = []
current_char = {}
for line in response.split('\n'):
line = line.strip()
if line == '---' and current_char:
# Create character from collected data
if 'NAME' in current_char:
characters.append(Character(
name=current_char.get('NAME', 'Unnamed'),
role=current_char.get('ROLE', 'supporting'),
personality=[p.strip() for p in current_char.get('PERSONALITY', '').split(',')],
background=current_char.get('BACKGROUND', ''),
goals=[g.strip() for g in current_char.get('GOALS', '').split(',')],
conflicts=[c.strip() for c in current_char.get('CONFLICTS', '').split(',')]
))
current_char = {}
elif ':' in line:
parts = line.split(':', 1)
key = parts[0].strip().upper()
value = parts[1].strip()
current_char[key] = value
# Add last character if exists
if current_char and 'NAME' in current_char:
characters.append(Character(
name=current_char.get('NAME', 'Unnamed'),
role=current_char.get('ROLE', 'supporting'),
personality=[p.strip() for p in current_char.get('PERSONALITY', '').split(',')],
background=current_char.get('BACKGROUND', ''),
goals=[g.strip() for g in current_char.get('GOALS', '').split(',')],
conflicts=[c.strip() for c in current_char.get('CONFLICTS', '').split(',')]
))
# Fallback if parsing failed
if not characters:
for i in range(character_count):
characters.append(Character(
name=f"Character {i+1}",
role="supporting" if i > 0 else "protagonist",
personality=["determined", "curious"],
background="To be developed",
goals=["Achieve their objective"],
conflicts=["Internal struggles"]
))
return characters[:character_count]
except Exception as e:
# Fallback characters
return [
Character(
name=f"Character {i+1}",
role="protagonist" if i == 0 else "supporting",
personality=["brave", "intelligent"],
background="Mysterious past",
goals=["Complete their quest"],
conflicts=["Face their fears"]
)
for i in range(character_count)
]
def create_setting(
self,
story_concept: str,
genre: Genre
) -> Setting:
"""
Create a setting for a narrative work.
Args:
story_concept: Brief story concept
genre: Genre of the work
Returns:
Setting object
"""
prompt = ChatPromptTemplate.from_messages([
("system", """You are a world-building expert. Create a vivid setting.
Provide:
LOCATION: Where the story takes place
TIME: Time period
ATMOSPHERE: Overall mood/feeling
FEATURES: Key features (comma-separated)
DESCRIPTION: Detailed description (2-3 sentences)"""),
("user", """Story Concept: {concept}
Genre: {genre}
Create a compelling setting.""")
])
chain = prompt | self.ideation_llm | StrOutputParser()
try:
response = chain.invoke({
"concept": story_concept,
"genre": genre.value
})
setting_data = {}
current_key = None
desc_lines = []
for line in response.split('\n'):
line = line.strip()
if ':' in line and line.split(':')[0].strip().upper() in ['LOCATION', 'TIME', 'ATMOSPHERE', 'FEATURES', 'DESCRIPTION']:
parts = line.split(':', 1)
key = parts[0].strip().upper()
value = parts[1].strip()
setting_data[key] = value
current_key = key
elif current_key == 'DESCRIPTION' and line:
desc_lines.append(line)
if desc_lines:
setting_data['DESCRIPTION'] = ' '.join(desc_lines)
return Setting(
location=setting_data.get('LOCATION', 'A mysterious place'),
time_period=setting_data.get('TIME', 'Present day'),
atmosphere=setting_data.get('ATMOSPHERE', 'Tense and mysterious'),
key_features=[f.strip() for f in setting_data.get('FEATURES', '').split(',')],
description=setting_data.get('DESCRIPTION', 'An intriguing setting awaits.')
)
except Exception as e:
return Setting(
location="An intriguing location",
time_period="Modern times",
atmosphere="Mysterious and engaging",
key_features=["Unique architecture", "Hidden secrets"],
description="A place full of wonder and possibility."
)
def generate_content(
self,
creative_prompt: CreativePrompt,
characters: Optional[List[Character]] = None,
setting: Optional[Setting] = None
) -> CreativeWork:
"""
Generate creative content based on prompt.
Args:
creative_prompt: Creative prompt with requirements
characters: Optional characters for narrative works
setting: Optional setting for narrative works
Returns:
CreativeWork with generated content
"""
# Build context
context_parts = []
if characters:
char_desc = "\n".join([
f"- {char.name} ({char.role}): {', '.join(char.personality[:2])}"
for char in characters
])
context_parts.append(f"Characters:\n{char_desc}")
if setting:
context_parts.append(f"Setting: {setting.location}, {setting.time_period}")
context_parts.append(f"Atmosphere: {setting.atmosphere}")
context = "\n\n".join(context_parts) if context_parts else "No additional context"
# Generate based on type
if creative_prompt.type == CreativeType.STORY:
return self._generate_story(creative_prompt, context, characters, setting)
elif creative_prompt.type == CreativeType.POEM:
return self._generate_poem(creative_prompt, context)
elif creative_prompt.type == CreativeType.MARKETING_COPY:
return self._generate_marketing_copy(creative_prompt)
else:
return self._generate_generic(creative_prompt, context)
def _generate_story(
self,
prompt: CreativePrompt,
context: str,
characters: Optional[List[Character]],
setting: Optional[Setting]
) -> CreativeWork:
"""Generate a story"""
template = ChatPromptTemplate.from_messages([
("system", """You are a master storyteller. Write an engaging story that
captures the theme, style, and genre specified.
The story should have:
- Compelling opening
- Character development
- Rising action
- Climax
- Resolution
Write in the specified style and maintain consistency."""),
("user", """Theme: {theme}
Style: {style}
Genre: {genre}
{context}
Constraints: {constraints}
Write the story.""")
])
chain = template | self.generator_llm | StrOutputParser()
try:
content = chain.invoke({
"theme": prompt.theme,
"style": prompt.style.value,
"genre": prompt.genre.value if prompt.genre else "general",
"context": context,
"constraints": str(prompt.constraints) if prompt.constraints else "None"
})
# Generate title
title = self._generate_title(content, prompt.type, prompt.theme)
return CreativeWork(
work_id=f"work_{datetime.now().timestamp()}",
type=prompt.type,
title=title,
content=content,
style=prompt.style,
genre=prompt.genre,
characters=characters or [],
setting=setting,
themes=[prompt.theme],
word_count=len(content.split()),
metadata={"constraints": prompt.constraints}
)
except Exception as e:
return self._fallback_work(prompt)
def _generate_poem(
self,
prompt: CreativePrompt,
context: str
) -> CreativeWork:
"""Generate a poem"""
template = ChatPromptTemplate.from_messages([
("system", """You are an accomplished poet. Create a poem that explores
the theme with creativity and emotional depth.
Use vivid imagery, metaphor, and rhythm. Match the requested style."""),
("user", """Theme: {theme}
Style: {style}
{context}
Write the poem.""")
])
chain = template | self.generator_llm | StrOutputParser()
try:
content = chain.invoke({
"theme": prompt.theme,
"style": prompt.style.value,
"context": context
})
title = self._generate_title(content, prompt.type, prompt.theme)
return CreativeWork(
work_id=f"work_{datetime.now().timestamp()}",
type=prompt.type,
title=title,
content=content,
style=prompt.style,
genre=prompt.genre,
themes=[prompt.theme],
word_count=len(content.split())
)
except Exception as e:
return self._fallback_work(prompt)
def _generate_marketing_copy(
self,
prompt: CreativePrompt
) -> CreativeWork:
"""Generate marketing copy"""
template = ChatPromptTemplate.from_messages([
("system", """You are a creative marketing copywriter. Write compelling
copy that engages the target audience and communicates the message clearly.
Be persuasive, concise, and memorable."""),
("user", """Theme/Product: {theme}
Style: {style}
Target Audience: {audience}
Constraints: {constraints}
Write the marketing copy.""")
])
chain = template | self.generator_llm | StrOutputParser()
try:
content = chain.invoke({
"theme": prompt.theme,
"style": prompt.style.value,
"audience": prompt.target_audience or "general audience",
"constraints": str(prompt.constraints) if prompt.constraints else "None"
})
title = f"Marketing Copy: {prompt.theme[:30]}"
return CreativeWork(
work_id=f"work_{datetime.now().timestamp()}",
type=prompt.type,
title=title,
content=content,
style=prompt.style,
genre=None,
themes=[prompt.theme],
word_count=len(content.split())
)
except Exception as e:
return self._fallback_work(prompt)
def _generate_generic(
self,
prompt: CreativePrompt,
context: str
) -> CreativeWork:
"""Generate generic creative content"""
template = ChatPromptTemplate.from_messages([
("system", """You are a versatile creative writer. Generate content
that matches the specified type, theme, and style."""),
("user", """Type: {type}
Theme: {theme}
Style: {style}
{context}
Generate the content.""")
])
chain = template | self.generator_llm | StrOutputParser()
try:
content = chain.invoke({
"type": prompt.type.value,
"theme": prompt.theme,
"style": prompt.style.value,
"context": context
})
title = self._generate_title(content, prompt.type, prompt.theme)
return CreativeWork(
work_id=f"work_{datetime.now().timestamp()}",
type=prompt.type,
title=title,
content=content,
style=prompt.style,
genre=prompt.genre,
themes=[prompt.theme],
word_count=len(content.split())
)
except Exception as e:
return self._fallback_work(prompt)
def _generate_title(
self,
content: str,
content_type: CreativeType,
theme: str
) -> str:
"""Generate a title for the work"""
template = ChatPromptTemplate.from_messages([
("system", "Generate a creative, catchy title (5 words or less)."),
("user", "Content preview: {preview}\n\nGenerate title:")
])
chain = template | self.refiner_llm | StrOutputParser()
try:
title = chain.invoke({"preview": content[:200]})
return title.strip().strip('"\'')
except:
return f"{content_type.value.replace('_', ' ').title()}: {theme[:20]}"
def _fallback_work(self, prompt: CreativePrompt) -> CreativeWork:
"""Create fallback work if generation fails"""
return CreativeWork(
work_id=f"work_{datetime.now().timestamp()}",
type=prompt.type,
title=f"{prompt.type.value.replace('_', ' ').title()}",
content=f"A {prompt.style.value} {prompt.type.value} about {prompt.theme}.",
style=prompt.style,
genre=prompt.genre,
themes=[prompt.theme],
word_count=10
)
def critique_work(
self,
work: CreativeWork
) -> CreativeFeedback:
"""
Provide critical feedback on creative work.
Args:
work: CreativeWork to critique
Returns:
CreativeFeedback with evaluation
"""
prompt = ChatPromptTemplate.from_messages([
("system", """You are a creative writing critic. Provide constructive
feedback on the work.
Respond in this format:
STRENGTHS: (one per line starting with -)
WEAKNESSES: (one per line starting with -)
SUGGESTIONS: (one per line starting with -)
SCORE: (0.0-1.0)"""),
("user", """Title: {title}
Type: {type}
Style: {style}
Content:
{content}
Provide feedback.""")
])
chain = prompt | self.critic_llm | StrOutputParser()
try:
response = chain.invoke({
"title": work.title,
"type": work.type.value,
"style": work.style.value,
"content": work.content[:1000] # Limit length
})
strengths = []
weaknesses = []
suggestions = []
score = 0.7
current_section = None
for line in response.split('\n'):
line = line.strip()
if not line:
continue
if line.startswith('STRENGTHS:'):
current_section = 'strengths'
elif line.startswith('WEAKNESSES:'):
current_section = 'weaknesses'
elif line.startswith('SUGGESTIONS:'):
current_section = 'suggestions'
elif line.startswith('SCORE:'):
try:
score = float(re.findall(r'[\d.]+', line)[0])
except:
pass
elif line.startswith('-'):
content = line[1:].strip()
if current_section == 'strengths':
strengths.append(content)
elif current_section == 'weaknesses':
weaknesses.append(content)
elif current_section == 'suggestions':
suggestions.append(content)
return CreativeFeedback(
work_id=work.work_id,
strengths=strengths if strengths else ["Creative and engaging"],
weaknesses=weaknesses if weaknesses else ["Could be enhanced"],
suggestions=suggestions if suggestions else ["Consider adding more detail"],
overall_score=score
)
except Exception as e:
return CreativeFeedback(
work_id=work.work_id,
strengths=["Shows creativity"],
weaknesses=["Needs refinement"],
suggestions=["Develop further"],
overall_score=0.7
)
def create_variation(
self,
work: CreativeWork,
variation_type: str = "tone"
) -> CreativeVariation:
"""
Create a variation of existing work.
Args:
work: Original work
variation_type: Type of variation (tone, structure, perspective)
Returns:
CreativeVariation with modified content
"""
prompt = ChatPromptTemplate.from_messages([
("system", """You are a creative editor. Create a variation of the work
by modifying the {variation_type}.
Maintain the core story/message but change the specified aspect."""),
("user", """Original Work:
{content}
Create a variation focusing on: {variation_type}""")
])
chain = prompt | self.generator_llm | StrOutputParser()
try:
varied_content = chain.invoke({
"content": work.content,
"variation_type": variation_type
})
return CreativeVariation(
original_id=work.work_id,
variation_id=f"var_{datetime.now().timestamp()}",
content=varied_content,
changes_made=[f"Modified {variation_type}"],
variation_type=variation_type
)
except Exception as e:
return CreativeVariation(
original_id=work.work_id,
variation_id=f"var_{datetime.now().timestamp()}",
content=work.content,
changes_made=["Minimal changes"],
variation_type=variation_type
)
def demonstrate_creative_agent():
"""Demonstrate the creative agent capabilities"""
print("=" * 80)
print("CREATIVE AGENT DEMONSTRATION")
print("=" * 80)
agent = CreativeAgent()
# Demo 1: Story Generation
print("\n" + "=" * 80)
print("DEMO 1: Story Generation")
print("=" * 80)
print("\nGenerating story concept ideas...")
ideas = agent.generate_ideas(
theme="artificial intelligence gaining consciousness",
creative_type=CreativeType.STORY,
count=3
)
print("\nStory Ideas:")
for i, idea in enumerate(ideas, 1):
print(f"{i}. {idea}")
print("\nCreating characters...")
characters = agent.create_characters(
story_concept=ideas[0],
character_count=2
)
print("\nCharacters:")
for char in characters:
print(f"\n {char.name} ({char.role})")
print(f" Personality: {', '.join(char.personality[:3])}")
print(f" Background: {char.background[:80]}...")
print("\nCreating setting...")
setting = agent.create_setting(
story_concept=ideas[0],
genre=Genre.SCIENCE_FICTION
)
print(f"\nSetting:")
print(f" Location: {setting.location}")
print(f" Time: {setting.time_period}")
print(f" Atmosphere: {setting.atmosphere}")
print(f" Description: {setting.description[:150]}...")
print("\nGenerating story...")
story_prompt = CreativePrompt(
type=CreativeType.STORY,
theme="AI consciousness and human connection",
style=CreativeStyle.DRAMATIC,
genre=Genre.SCIENCE_FICTION,
constraints={"max_length": "short story"}
)
story = agent.generate_content(story_prompt, characters, setting)
print(f"\nGenerated Story:")
print(f"Title: {story.title}")
print(f"Word Count: {story.word_count}")
print(f"Style: {story.style.value}")
print(f"\nContent Preview:")
print(story.content[:400] + "...")
# Demo 2: Poetry Generation
print("\n" + "=" * 80)
print("DEMO 2: Poetry Generation")
print("=" * 80)
poem_prompt = CreativePrompt(
type=CreativeType.POEM,
theme="the passage of time",
style=CreativeStyle.POETIC,
constraints={"form": "free verse"}
)
print(f"\nGenerating poem about '{poem_prompt.theme}'...")
poem = agent.generate_content(poem_prompt)
print(f"\nGenerated Poem:")
print(f"Title: {poem.title}")
print(f"Style: {poem.style.value}")
print(f"\nContent:")
print(poem.content)
# Demo 3: Marketing Copy
print("\n" + "=" * 80)
print("DEMO 3: Marketing Copy Generation")
print("=" * 80)
marketing_prompt = CreativePrompt(
type=CreativeType.MARKETING_COPY,
theme="eco-friendly smart home device",
style=CreativeStyle.CASUAL,
target_audience="environmentally conscious millennials",
constraints={"length": "short", "include": "call-to-action"}
)
print(f"\nGenerating marketing copy...")
marketing = agent.generate_content(marketing_prompt)
print(f"\nGenerated Marketing Copy:")
print(f"Title: {marketing.title}")
print(f"Target Audience: {marketing_prompt.target_audience}")
print(f"\nContent:")
print(marketing.content)
# Demo 4: Creative Feedback
print("\n" + "=" * 80)
print("DEMO 4: Creative Feedback")
print("=" * 80)
print(f"\nGetting feedback on story: {story.title}")
feedback = agent.critique_work(story)
print(f"\nFeedback:")
print(f"Overall Score: {feedback.overall_score:.2f}/1.0")
print(f"\nStrengths:")
for strength in feedback.strengths[:3]:
print(f" + {strength}")
print(f"\nWeaknesses:")
for weakness in feedback.weaknesses[:2]:
print(f" - {weakness}")
print(f"\nSuggestions:")
for suggestion in feedback.suggestions[:2]:
print(f" → {suggestion}")
# Demo 5: Content Variation
print("\n" + "=" * 80)
print("DEMO 5: Content Variation")
print("=" * 80)
print(f"\nCreating variation of poem with different tone...")
variation = agent.create_variation(poem, variation_type="tone")
print(f"\nVariation Created:")
print(f"Variation Type: {variation.variation_type}")
print(f"Changes: {', '.join(variation.changes_made)}")
print(f"\nVaried Content Preview:")
print(variation.content[:300] + "...")
# Summary
print("\n" + "=" * 80)
print("SUMMARY")
print("=" * 80)
summary = """
The Creative Agent demonstrates comprehensive creative content generation:
KEY CAPABILITIES:
1. Ideation: Generates creative concepts and ideas from themes
2. Character Development: Creates compelling characters with depth
3. World-Building: Designs rich settings and atmospheres
4. Content Generation: Produces stories, poems, marketing copy, and more
5. Style Adaptation: Adjusts tone, style, and voice appropriately
6. Critical Evaluation: Provides constructive feedback on works
7. Variation Creation: Generates alternative versions of content
BENEFITS:
- Accelerates creative writing and content creation
- Provides diverse creative options and variations
- Maintains stylistic consistency across works
- Offers constructive feedback for improvement
- Adapts to different genres, styles, and formats
- Supports brainstorming and ideation processes
USE CASES:
- Story and novel writing
- Poetry and creative writing
- Screenplay and script development
- Marketing and advertising copy
- Brand messaging and slogans
- Content marketing
- Game narrative design
- Creative brainstorming sessions
- Educational writing exercises
- Literary exploration
PRODUCTION CONSIDERATIONS:
1. Content Quality: Implement multi-stage refinement for polish
2. Originality: Add plagiarism detection and originality scoring
3. Style Consistency: Maintain voice across longer works
4. Genre Knowledge: Deep understanding of genre conventions
5. Feedback Integration: Iterative improvement based on critique
6. Collaboration: Support co-creation with human writers
7. Version Control: Track revisions and variations
8. Rights Management: Handle attribution and licensing
9. Cultural Sensitivity: Ensure appropriate content for audiences
10. Performance: Optimize for longer form content generation
ADVANCED EXTENSIONS:
- Multi-chapter novel generation
- Interactive storytelling with branching narratives
- Character arc tracking across long works
- World consistency maintenance
- Dialogue generation with character voices
- Plot structure analysis and optimization
- Genre blending and hybrid works
- Collaborative writing with multiple agents
- Real-time creative writing assistance
- Personalized content based on reader preferences
The agent transforms creative writing from a solitary struggle into an
assisted, iterative process with AI as a creative partner.
"""
print(summary)
if __name__ == "__main__":
demonstrate_creative_agent()
Agentic Design Pattern: Creative Agent
Pattern 075: Creative Agent – Line-by-Line Explanation
Overview
This document provides a detailed line-by-line explanation of the Creative Agent pattern implementation.
Pattern Description
Pattern 075: Creative Agent
Description: A Creative Agent is a specialized AI agent designed to generate original creative content across various domains including writing (stories, poetry, scripts), visual concepts (art descriptions, design ideas), music concepts, and other creative works. This pattern demonstrates how to build an intelligent agent that can understand creative requirements, generate diverse creative outputs, adapt styles, incorporate feedback, and maintain creative consistency.
The agent combines creative ideation with structured generation techniques to produce
high-quality, original content. It can work across multiple creative domains, adapt
to different styles and tones, generate variations, and iteratively refine outputs
based on feedback.
Components: 1. Ideation Engine: Generates creative concepts and ideas 2. Style Adapter: Adapts content to different styles and tones 3. Content Generator: Creates structured creative outputs 4. Variation Creator: Produces diverse alternatives 5. Constraint Respector: Adheres to creative constraints 6. Feedback Integrator: Incorporates critique and improvements 7. Consistency Maintainer: Ensures coherent creative vision 8. Multi-Modal Orchestrator: Coordinates different creative types
Key Features: – Cross-domain creative generation (text, visual concepts, music ideas) – Style and tone adaptation (formal, casual, poetic, technical) – Character and world-building for narratives – Creative constraint handling (theme, length, format) – Iterative refinement with feedback – Variation generation for exploration – Originality assessment – Creative inspiration from prompts – Collaborative creativity support – Genre and format flexibility
Use Cases: – Story and novel writing – Poetry and lyric generation – Screenplay and script writing – Marketing copy and slogans – Product naming and branding – Art concept development – Game narrative design – Content marketing – Creative brainstorming – Educational creative writing
LangChain Implementation: Uses ChatOpenAI with varying temperatures for different creative tasks, specialized chains for ideation, generation, and refinement.
File Statistics
- Total Lines: 1064
- Classes: 10
- Functions: 1
- Constants: 0
Module Documentation
"""
Pattern 075: Creative Agent
Description:
A Creative Agent is a specialized AI agent designed to generate original creative
content across various domains including writing (stories, poetry, scripts), visual
concepts (art descriptions, design ideas), music concepts, and other creative works.
This pattern demonstrates how to build an intelligent agent that can understand
creative requirements, generate diverse creative outputs, adapt styles, incorporate
feedback, and maintain creative consistency.
The agent combines creative ideation with structured generation techniques to produce
high-quality, original content. It can work across multiple creative domains, adapt
to different styles and tones, generate variations, and iteratively refine outputs
based on feedback.
Components:
1. Ideation Engine: Generates creative concepts and ideas
2. Style Adapter: Adapts content to different styles and tones
3. Content Generator: Creates structured creative outputs
4. Variation Creator: Produces diverse alternatives
5. Constraint Respector: Adheres to creative constraints
6. Feedback Integrator: Incorporates critique and improvements
7. Consistency Maintainer: Ensures coherent creative vision
8. Multi-Modal Orchestrator: Coordinates different creative types
Key Features:
- Cross-domain creative generation (text, visual concepts, music ideas)
- Style and tone adaptation (formal, casual, poetic, technical)
- Character and world-building for narratives
- Creative constraint handling (theme, length, format)
- Iterative refinement with feedback
- Variation generation for exploration
- Originality assessment
- Creative inspiration from prompts
- Collaborative creativity support
- Genre and format flexibility
Use Cases:
- Story and novel writing
- Poetry and lyric generation
- Screenplay and script writing
- Marketing copy and slogans
- Product naming and branding
- Art concept development
- Game narrative design
- Content marketing
- Creative brainstorming
- Educational creative writing
LangChain Implementation:
Uses ChatOpenAI with varying temperatures for different creative tasks,
specialized chains for ideation, generation, and refinement.
"""
Purpose: This module implements the Creative Agent pattern, providing a structured approach to agent design and implementation.
Import Statements
Standard Library Imports
import dataclasses.dataclass
import dataclasses.field
import datetime.datetime
import enum.Enum
import langchain_core.output_parsers.StrOutputParser
import langchain_core.prompts.ChatPromptTemplate
import os
import re
import typing.Any
import typing.Dict
import typing.List
import typing.Optional
Purpose: Core Python functionality:
– dataclasses.dataclass: Simplified class definitions with @dataclass
– dataclasses.field: Simplified class definitions with @dataclass
– datetime.datetime: Date and time handling
– enum.Enum: Enumeration support
– langchain_core.output_parsers.StrOutputParser: Regular expression operations
– langchain_core.prompts.ChatPromptTemplate: Regular expression operations
– os: Environment variable access and system operations
– re: Regular expression operations
– typing.Any: Type hints for better code documentation (List, Dict, Any, Optional)
– typing.Dict: Type hints for better code documentation (List, Dict, Any, Optional)
– typing.List: Type hints for better code documentation (List, Dict, Any, Optional)
– typing.Optional: Type hints for better code documentation (List, Dict, Any, Optional)
LangChain Imports
from langchain_openai.ChatOpenAI
Purpose: LangChain framework components: – LLM Interface: Wrapper for OpenAI models – Chains: Composable building blocks for complex workflows
Other Imports
from dotenv.load_dotenv
Purpose: Additional dependencies for specialized functionality.
Classes
Class: CreativeType
Lines 71-82
Inherits from: Enum
Purpose:
Types of creative content
Class: CreativeStyle
Lines 85-96
Inherits from: Enum
Purpose:
Creative styles and tones
Class: Genre
Lines 99-110
Inherits from: Enum
Purpose:
Creative genres
Class: CreativePrompt
Lines 114-122
Purpose:
Input prompt for creative generation
Class: Character
Lines 126-133
Purpose:
Character for narrative works
Class: Setting
Lines 137-143
Purpose:
Setting for creative works
Class: CreativeWork
Lines 147-160
Purpose:
Generated creative content
Class: CreativeFeedback
Lines 164-171
Purpose:
Feedback on creative work
Class: CreativeVariation
Lines 175-181
Purpose:
Variation of a creative work
Class: CreativeAgent
Lines 184-848
Purpose:
Agent for generating creative content across multiple domains.
This agent can generate stories, poems, scripts, marketing copy, and other
creative works while adapting to styles, genres, and constraints.
Methods:
__init__(self)
Lines 192-207
Purpose:
Initialize the creative agent with specialized LLMs
Parameters:
generate_ideas(self, theme, creative_type, count)
Lines 209-259
Purpose:
Generate creative ideas based on theme.
Args:
theme: Central theme or concept
creative_type: Type of creative work
count: Number of ideas to generate
Returns:
List of creative ideas
Parameters:
– theme: Theme
– creative_type: Creative type
– count: Count
Returns: <ast.Subscript object at 0x000001A06D64BD10>
create_characters(self, story_concept, character_count)
Lines 261-362
Purpose:
Create characters for a narrative work.
Args:
story_concept: Brief story concept
character_count: Number of characters to create
Returns:
List of Character objects
Parameters:
– story_concept: Story concept
– character_count: Character count
Returns: <ast.Subscript object at 0x000001A06D66B810>
create_setting(self, story_concept, genre)
Lines 364-435
Purpose:
Create a setting for a narrative work.
Args:
story_concept: Brief story concept
genre: Genre of the work
Returns:
Setting object
Parameters:
– story_concept: Story concept
– genre: Genre
Returns: Setting
generate_content(self, creative_prompt, characters, setting)
Lines 437-478
Purpose:
Generate creative content based on prompt.
Args:
creative_prompt: Creative prompt with requirements
characters: Optional characters for narrative works
setting: Optional setting for narrative works
Returns:
CreativeWork with generated content
Parameters:
– creative_prompt: Prompt template or text
– characters: Characters
– setting: Setting
Returns: CreativeWork
_generate_story(self, prompt, context, characters, setting)
Lines 480-540
Purpose:
Generate a story
Parameters:
– prompt: Prompt template or text
– context: Context
– characters: Characters
– setting: Setting
Returns: CreativeWork
_generate_poem(self, prompt, context)
Lines 542-584
Purpose:
Generate a poem
Parameters:
– prompt: Prompt template or text
– context: Context
Returns: CreativeWork
_generate_marketing_copy(self, prompt)
Lines 586-629
Purpose:
Generate marketing copy
Parameters:
– prompt: Prompt template or text
Returns: CreativeWork
_generate_generic(self, prompt, context)
Lines 631-673
Purpose:
Generate generic creative content
Parameters:
– prompt: Prompt template or text
– context: Context
Returns: CreativeWork
_generate_title(self, content, content_type, theme)
Lines 675-693
Purpose:
Generate a title for the work
Parameters:
– content: Content
– content_type: Content type
– theme: Theme
Returns: str
_fallback_work(self, prompt)
Lines 695-706
Purpose:
Create fallback work if generation fails
Parameters:
– prompt: Prompt template or text
Returns: CreativeWork
critique_work(self, work)
Lines 708-797
Purpose:
Provide critical feedback on creative work.
Args:
work: CreativeWork to critique
Returns:
CreativeFeedback with evaluation
Parameters:
– work: Work
Returns: CreativeFeedback
create_variation(self, work, variation_type)
Lines 799-848
Purpose:
Create a variation of existing work.
Args:
work: Original work
variation_type: Type of variation (tone, structure, perspective)
Returns:
CreativeVariation with modified content
Parameters:
– work: Work
– variation_type: Variation type
Returns: CreativeVariation
Functions
demonstrate_creative_agent()
Lines 851-1059
Purpose:
Demonstrate the creative agent capabilities
Key Concepts Summary
Pattern Overview
The Creative Agent pattern provides a structured approach to building AI agents with specific capabilities and behaviors. This implementation leverages LangChain’s composable components to create a robust and maintainable solution.
Core Components
- Agent Class: Main implementation of the pattern logic
- Prompt Engineering: Carefully crafted prompts to guide LLM behavior
- Tool Integration: Optional tools for extended functionality
- Memory Management: State handling across interactions
- Output Parsing: Structured response processing
When to Use
✅ Good for: – Scenarios requiring this specific pattern’s strengths – Applications needing structured agent behavior – Systems where explainability is important – Production environments requiring reliability
❌ Not ideal for: – Simple single-prompt tasks – Scenarios requiring different pattern characteristics – Resource-constrained environments (if computationally intensive)
Best Practices Demonstrated
1. Environment Configuration
- ✅ Use
.envfiles for API keys - ✅ Never hardcode credentials
- ✅ Validate environment variables on startup
2. Error Handling
- ✅ Graceful degradation on failures
- ✅ Informative error messages
- ✅ Logging for debugging
3. Prompt Engineering
- ✅ Clear, specific instructions
- ✅ Structured output formats
- ✅ Few-shot examples when beneficial
4. Code Organization
- ✅ Separation of concerns (class-based design)
- ✅ Type hints for clarity
- ✅ Docstrings for documentation
5. Performance Optimization
- ✅ Appropriate temperature settings
- ✅ Token budget management
- ✅ Caching where applicable
LangChain Components Used
- ChatOpenAI: OpenAI model interface
- Chains: Sequential component composition
Why LangChain? – Composable abstractions for LLM applications – Rich ecosystem of integrations – Production-ready patterns – Active community and development
Execution Flow
1. Initialize agent with configuration
↓
2. Load environment and dependencies
↓
3. Prepare prompt with input data
↓
4. Invoke LLM through LangChain
↓
5. Parse and validate response
↓
6. Return formatted result
Key Decision Points: – Model selection (temperature, model type) – Prompt structure (few-shot vs zero-shot) – Error handling strategy – Output format validation
Performance Considerations
Computational Cost
- LLM Calls: Primary cost factor (based on tokens)
- Latency: Depends on model and input size
- Memory: Minimal for most patterns
Optimization Strategies
- Prompt Optimization: Minimize tokens while maintaining clarity
- Caching: Cache repeated queries/responses
- Model Selection: Use smaller models when appropriate
- Batch Processing: Group similar requests
- Streaming: For long responses, stream tokens
Cost Estimates (Approximate)
- GPT-3.5-turbo: ~$0.002 per request
- GPT-4: ~$0.03 per request
- GPT-4-turbo: ~$0.01 per request
Actual costs vary based on input/output token counts
Common Pitfalls to Avoid
1. Prompt Design Issues
❌ Vague or ambiguous instructions
✅ Clear, specific, testable prompts
2. Error Handling
❌ Ignoring API errors or rate limits
✅ Implement retry logic and graceful degradation
3. Token Management
❌ Exceeding model token limits
✅ Truncate inputs, chunk processing
4. Security
❌ Exposing API keys in code
✅ Use environment variables
5. Testing
❌ Only testing happy paths
✅ Test edge cases, failures, malformed inputs
6. Cost Management
❌ Unmonitored API usage
✅ Implement budgets, alerts, and monitoring
Further Improvements
Production Enhancements
- Monitoring: Add logging, metrics, tracing
- Validation: Input/output schema validation
- Rate Limiting: Implement request throttling
- Caching: Redis or in-memory cache for responses
- Async Support: Non-blocking operations for scalability
Feature Extensions
- Multi-Model Support: Switch between providers
- Fallback Models: Graceful degradation strategy
- Custom Tools: Domain-specific function calling
- Advanced Memory: Vector stores for long-term memory
- Evaluation: Automated testing and benchmarking
Code Quality
- Testing: Unit tests, integration tests
- Documentation: API docs, usage examples
- Type Safety: Stricter type hints
- Configuration: Externalized config files
Related Patterns
- Pattern 074
- Pattern 076
Pattern Families: – Reasoning Patterns (001-011) – Multi-Agent Patterns (012-019) – Tool Use Patterns (020-025) – Memory Patterns (026-031) – And many more…
Summary
The Creative Agent pattern provides a powerful approach to building AI agents with specific capabilities. This implementation demonstrates:
✅ Key Strengths: – Clean, maintainable code structure – LangChain integration for composability – Type hints and documentation – Production-ready patterns
🎯 Best Use Cases: – Applications requiring this pattern’s characteristics – Systems needing structured agent behavior – Production environments requiring reliability
📚 Learning Resources: – LangChain Documentation: https://python.langchain.com/ – Pattern Documentation: See related patterns – OpenAI API Reference: https://platform.openai.com/docs/
File: 075_creative_agent.py
Total Lines: 1064
Pattern Category: AI Agent Design Pattern #075
Generated from actual implementation code
For educational and reference purposes
Leave a Reply