Quick Start
ActiveGenie delivers on three core promises: Real Value through intelligent AI analysis, Consistent results across providers and models, and Model-Agnostic flexibility to use any AI service.
Extractor
Transform unstructured text into reliable structured data using AI-powered analysis. Extract typed data from messy real-world content while handling informal language, rhetorical devices, and conversational patterns.
# Extract from product listing with inconsistent formatting
product_text = "Sony 65\" BRAVIA XR - $1999.99 (Save $500!) Amazing 4K HDR quality"
schema = {
brand: { type: 'string', description: 'Product brand' },
display_size: { type: 'string', description: 'Screen size with units' },
price: { type: 'number', minimum: 0, description: 'Current price' },
discount: { type: 'number', minimum: 0, description: 'Discount amount' },
features: { type: 'array', items: { type: 'string' }, description: 'Key features' }
}
# Works consistently across any AI provider
result = ActiveGenie::Extractor.call(product_text, schema,
config: { provider_name: :openai, model: 'gpt-4o-mini' })
# => {
# brand: "Sony",
# display_size: "65\"",
# price: 1999.99,
# discount: 500.0,
# features: ["BRAVIA XR", "4K HDR quality"]
# }
# Handle informal language and rhetorical devices
review = "The new update isn't terrible, but it's not exactly amazing either"
sentiment_schema = {
sentiment: { type: 'string', enum: ['positive', 'negative', 'neutral', 'mixed'] },
satisfaction_level: { type: 'integer', minimum: 1, maximum: 5 }
}
result = ActiveGenie::Extractor.with_litote(review, sentiment_schema)
# => {
# sentiment: "mixed",
# satisfaction_level: 3,
# message_litote: true,
# litote_rephrased: "The new update is okay, but it could be better"
# }Scorer
Objective evaluation system using AI-powered expert reviewers. Get detailed scoring with transparent reasoning from multiple domain experts that automatically adapt to your content.
# Automatic expert selection based on content
code_review = "Added rate limiting with sliding window algorithm, includes comprehensive unit tests and performance benchmarks"
criteria = "Evaluate technical quality, completeness, and engineering best practices"
# Same interface works with any AI provider
result = ActiveGenie::Scorer.call(code_review, criteria,
config: { provider_name: :anthropic, model: 'claude-3-5-haiku' })
# => {
# senior_software_engineer_score: 94,
# senior_software_engineer_reasoning: "Excellent implementation with proper algorithm choice, comprehensive testing, and performance considerations",
# technical_architect_score: 89,
# technical_architect_reasoning: "Strong technical approach, could benefit from documentation of rate limiting strategy",
# devops_engineer_score: 91,
# devops_engineer_reasoning: "Good performance benchmarking approach, suggests monitoring and observability awareness",
# final_score: 91.3
# }
# Custom expert reviewers for specialized evaluation
medical_text = "Patient shows 17% improvement in cardiac ejection fraction following 6-week therapy protocol"
reviewers = ["Cardiologist", "Clinical Researcher", "Medical Writer"]
result = ActiveGenie::Scorer.call(medical_text, "Evaluate clinical accuracy and reporting quality", reviewers)Comparator
Structured debate system that determines winners through AI-powered analysis. Conducts verbal debates where players present arguments, counter-arguments, and final statements before an impartial judge decides.
# Code quality comparison with detailed reasoning
player_a = "Implementation uses dependency injection with comprehensive interfaces, enabling easy testing and component replacement"
player_b = "Code achieves 95% test coverage using traditional patterns with proven stability and team familiarity"
criteria = "Evaluate long-term maintainability, testability, and team productivity"
# Consistent debate structure across all AI providers
result = ActiveGenie::Comparator.call(player_a, player_b, criteria,
config: { provider_name: :google, model: 'gemini-2.0-flash' })
# => {
# winner: "Implementation uses dependency injection...",
# loser: "Code achieves 95% test coverage...",
# reasoning: "Player A's dependency injection approach provides superior long-term maintainability through loose coupling, while Player B's high coverage is valuable but doesn't address the structural concerns for future development"
# }
# Specialized fight mode for character battles
fighter_a = "Master Crane: graceful fighter using Crane Kung Fu with lightness, precision, and momentum redirection"
fighter_b = "Iron Ox: powerful brawler using Ox Bull Charge style with immense strength and overwhelming mass"
fight_criteria = "Determine winner in one-on-one duel based on skill, strategy, and adaptability"
result = ActiveGenie::Comparator.by_fight(fighter_a, fighter_b, fight_criteria)Real Value: Structured debate format provides comprehensive analysis of complex comparisons
Consistent: Same debate process and winner determination logic across all AI providers
Model-Agnostic: Debate structure adapts to different AI reasoning capabilities seamlessly
Recommended models: claude-3-5-sonnet, gpt-4, gemini-2.0-flash
Ranker
Sophisticated multi-stage ranking system combining ActiveGenie::Scorer + ActiveGenie::Comparator, statistical elimination, ELO algorithms, and head-to-head battles to produce fair and accurate rankings for any number of players.
# Rank API approaches for a specific use case
api_options = [
"REST API with comprehensive OpenAPI documentation and versioning strategy",
"GraphQL API with efficient query resolution and real-time subscriptions",
"gRPC API with Protocol Buffers and bi-directional streaming capabilities",
"WebSocket API with custom protocol and connection state management"
]
criteria = "Best choice for real-time collaborative application with complex data relationships"
# Automatic methodology selection based on player count
result = ActiveGenie::Ranker.call(api_options, criteria,
config: { provider_name: :openai, model: 'gpt-4o' })
# => {
# players: [
# { content: "GraphQL API with efficient query...", score: 89, elo: 1245, rank: 1 },
# { content: "WebSocket API with custom protocol...", score: 85, elo: 1198, rank: 2 },
# { content: "gRPC API with Protocol Buffers...", score: 78, elo: 1156, rank: 3 },
# { content: "REST API with comprehensive...", score: 72, elo: 1089, rank: 4 }
# ],
# statistics: { total_players: 4, elo_rounds: 2, ffa_matches: 6 }
# }
# Tournament mode for comprehensive ranking
result = ActiveGenie::Ranker.by_tournament(api_options, criteria)
# ELO-only ranking for competitive scenarios
result = ActiveGenie::Ranker.by_elo(api_options, criteria)
# Simple scoring without battles
result = ActiveGenie::Ranker.by_scoring(api_options, criteria)Real Value: Multi-stage process eliminates inconsistent performers and produces statistically sound rankings
Consistent: Same ranking methodology and fairness algorithms across all AI providers
Model-Agnostic: Ranking logic works with any AI model's scoring and comparison capabilities
Recommended models: gpt-4, claude-3-5-sonnet, gemini-2.0-flash
Lister
Real Value + Consistent + Model-Agnostic
Generate ordered lists based on themes using "Family Feud" style survey simulation. Produces lists that reflect general public opinion and cultural consensus, perfect for market research and content planning.
# Market research for product development
theme = "Features smartphone users care about most when choosing a new device"
# Works identically across all AI providers
result = ActiveGenie::Lister.call(theme,
config: { provider_name: :anthropic, model: 'claude-3-5-haiku' })
# => [
# "Battery life",
# "Camera quality",
# "Price",
# "Storage capacity",
# "Brand reputation",
# "Screen size",
# "Processing speed"
# ]
# Generate expert jury recommendations
content = "Technical proposal for implementing microservices architecture with event-driven communication"
evaluation_criteria = "Assess technical feasibility, business impact, and implementation complexity"
experts = ActiveGenie::Lister.with_juries(content, evaluation_criteria)
# => [
# "Software Architect",
# "DevOps Engineer",
# "Business Analyst",
# "Technical Product Manager"
# ]
# Custom list size
result = ActiveGenie::Lister.call("Most popular breakfast foods worldwide",
config: { number_of_items: 10 })Real Value: Survey-style methodology produces culturally accurate and representative results
Consistent: Same ordering logic and popularity assessment across all AI providers
Model-Agnostic: List generation adapts to different AI models' cultural knowledge and reasoning
Recommended models: gpt-4o-mini, claude-3-5-haiku, gemini-2.0-flash
Model-Agnostic Configuration
Switch between any AI provider without changing your application logic:
# OpenAI
config = { provider_name: :openai, model: 'gpt-4o-mini' }
# Anthropic
config = { provider_name: :anthropic, model: 'claude-3-5-haiku' }
# Google
config = { provider_name: :google, model: 'gemini-2.0-flash' }
# DeepSeek
config = { provider_name: :deepseek, model: 'deepseek-chat' }
# Use the same interface for any module
ActiveGenie::Extractor.call(text, schema, config: config)
ActiveGenie::Scorer.call(text, criteria, config: config)
ActiveGenie::Comparator.call(player_a, player_b, criteria, config: config)
ActiveGenie::Ranker.call(players, criteria, config: config)
ActiveGenie::Lister.call(theme, config: config)Learn More
- Extractor Module - Advanced schemas, informal text processing, and detailed extraction patterns
- Scorer Module - Custom reviewers, scoring methodologies, and evaluation best practices
- Comparator Module - Debate structures, fight mode, and comparison strategies
- Ranker Module - Tournament systems, ELO algorithms, and ranking configurations
- Lister Module - Survey methodologies, expert juries, and list generation techniques