rubyllm-schema

name	rubyllm/schema
version	0.3.0
description	Ruby DSL for JSON Schema creation. Use this skill when defining structured data schemas for LLM function calling or structured outputs with RubyLLM. Provides Rails-inspired API for creating complex nested schemas.

RubyLLM::Schema v{{ page.version }}

Ruby DSL for JSON Schema Creation

A clean, Rails-inspired API for creating JSON schemas. Perfect for structured data schemas for LLM function calling or structured outputs.

Gem Version: 0.3.0
GitHub: https://github.com/danielfriis/ruby_llm-schema

Installation

gem 'ruby_llm-schema'
# or
bundle add ruby_llm-schema

Basic Usage

class PersonSchema < RubyLLM::Schema
  string :name, description: "Person's full name"
  integer :age, description: "Age in years", minimum: 0, maximum: 120
  boolean :active, required: false
  
  object :address do
    string :street
    string :city
    string :country, required: false
  end
  
  array :tags, of: :string, description: "User tags"
  
  array :contacts do
    object do
      string :email, format: "email"
      string :phone, required: false
    end
  end
  
  any_of :status do
    string enum: ["active", "pending", "inactive"]
    null
  end
end

# Generate JSON Schema
puts PersonSchema.to_json_schema

With RubyLLM

class PersonSchema < RubyLLM::Schema
  string :name, description: "Person's full name"
  integer :age, description: "Person's age in years"
  string :city, required: false, description: "City where they live"
end

# Use natively with RubyLLM
chat = RubyLLM.chat
response = chat.with_schema(PersonSchema)
  .ask("Generate a person named Alice who is 30 years old")

# Response is automatically parsed
puts response.content # => {"name" => "Alice", "age" => 30}
puts response.content.class # => Hash

Type Helpers

Primitive Types

string :name, description: "Full name"
integer :age, minimum: 0, maximum: 120
number :price, minimum: 0.0
boolean :active, default: true
null :optional_field

Complex Types

# Object
object :address do
  string :street
  string :city
  string :country
end

# Array of primitives
array :tags, of: :string
array :scores, of: :number

# Array of objects
array :contacts do
  object do
    string :email, format: "email"
    string :phone
  end
end

# Union types
any_of :value do
  string
  integer
  null
end

Enums

string :status, enum: ["active", "pending", "inactive"]

any_of :priority do
  string enum: ["low", "medium", "high"]
  null
end

Options

All type helpers accept these options:

• description - Schema description
• required - Whether field is required (default: true)
• default - Default value
• enum - Array of allowed values
• format - String format (email, uri, date-time, etc.)
• minimum/maximum - For numbers
• min_length/max_length - For strings
• items - For arrays

string :email, 
  description: "User email address",
  format: "email",
  min_length: 5,
  max_length: 255

number :price,
  minimum: 0,
  maximum: 10000,
  default: 0.0

Advanced Features

Schema Definitions

class ProductSchema < RubyLLM::Schema
  # Define reusable schema components
  definition :money do
    object do
      number :amount
      string :currency, enum: ["USD", "EUR", "GBP"]
    end
  end
  
  string :name
  ref :price, definition: :money
end

Inheritance

class BaseSchema < RubyLLM::Schema
  string :id
  timestamp :created_at
end

class UserSchema < BaseSchema
  string :name
  string :email
end

Strict Mode

class StrictSchema < RubyLLM::Schema
  strict true  # additionalProperties: false
  
  string :name
  # No other properties allowed
end

Use Cases

Article Metadata Extraction

class ArticleSchema < RubyLLM::Schema
  string :title
  array :topics, of: :string
  string :summary, description: "2-3 sentence summary"
  string :sentiment, enum: ["positive", "neutral", "negative"]
end

chat.with_schema(ArticleSchema)
  .ask("Extract metadata from this article: #{content}")

Customer Feedback Analysis

class FeedbackSchema < RubyLLM::Schema
  string :category, enum: ["bug", "feature", "usability", "other"]
  number :urgency, minimum: 1, maximum: 5
  string :summary
  array :action_items, of: :string
end

Structured Actions

class ActionSchema < RubyLLM::Schema
  string :action_type, enum: ["email", "call", "meeting", "task"]
  string :assignee
  timestamp :due_date
  object :context do
    string :priority
    array :dependencies, of: :string
  end
end

JSON Output

schema = PersonSchema.new
puts schema.to_json
# => {
#   "type": "object",
#   "properties": {
#     "name": { "type": "string", "description": "..." },
#     "age": { "type": "integer", "minimum": 0, "maximum": 120 }
#   },
#   "required": ["name", "age"]
# }

See Also

• Main RubyLLM: rubyllm
• Tools: tools
• Source: ~/Code/kaka/skills/.repos/ruby_llm-schema