ActiveGenie 🧞‍♂️

The Lodash for GenAI: Consistent + Model-Agnostic

ActiveGenie is a toolkit for building GenAI features in Ruby. Its modules are model-agnostic and work across providers. ActiveGenie::Comparator settles subjective comparisons by staging a structured debate between the two inputs. ActiveGenie::Scorer grades content with a jury of domain experts. ActiveGenie::Ranker orders large datasets with a tournament system.

Results stay consistent for three reasons:

• Custom benchmarking: every version and model update is tested for consistency.
• Reasoning prompting: prompts borrow human reasoning techniques, like debate and jury review, to shape how the model thinks.
• Overfitted prompts: each module uses a specialized prompt built for one purpose.

• Ruby >= 3.4.0
• An API key for OpenAI, Anthropic, DeepSeek, or Google

gem 'active_genie'

bundle install

Export a key for any supported provider:

export OPENAI_API_KEY="sk-..."

See the installation guide for Rails setup and explicit configuration.

require 'active_genie'

text = "Nike Air Max 90 - Size 42 - $199.99"
schema = {
  brand: { type: 'string', enum: %w[Nike Adidas Puma] },
  price: { type: 'number', minimum: 0 },
  size:  { type: 'number', minimum: 35, maximum: 46 }
}

result = ActiveGenie::Extractor.call(text, schema)

result.data
# => { brand: "Nike", price: 199.99, size: 42 }

result.reasoning
# => "Brand name appears at the start of the listing, price in USD at the end."

Every module returns an ActiveGenie::Result with data, reasoning, and metadata.

ActiveGenie::Scorer.call(text, criteria).data       # => 91
ActiveGenie::Comparator.call(a, b, criteria).data   # => the winning input
ActiveGenie::Lister.call(theme).data                # => ["Price", "Battery life", ...]
ActiveGenie::Ranker.call(items, criteria).data      # => items, best first

Full documentation is at activegenie.ai:

• Quickstart
• Configuration
• Benchmark results

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

This project is licensed under the Apache License 2.0. See the LICENSE file for details.