AI Coding with Ruby

AI Tool Ecosystem for Ruby

Ruby's AI tool ecosystem is solid, buoyed by Rails' dominance and the massive amount of open-source Ruby code on GitHub. GitHub Copilot produces very good Rails code because Rails' convention-over-configuration philosophy creates predictable patterns. Cursor and Claude Code understand ActiveRecord associations, concerns, and the Rails directory structure well. However, Ruby's dynamic metaprogramming means AI tools sometimes cannot trace method definitions at generation time, reducing accuracy for heavily meta-programmed codebases. Sorbet adoption is growing and when present, significantly improves AI tool accuracy. RuboCop integration provides style feedback on AI output. The ecosystem is notably weaker for non-Rails Ruby -- AI tools have much less training data for Sinatra, Hanami, or standalone Ruby scripts.

What AI Does Well with Ruby

• Generating complete Rails MVC layers (model, controller, views, routes) following naming conventions automatically
• Writing RSpec test suites with proper describe/context nesting, let blocks, and shared examples
• Creating ActiveRecord migrations, associations, scopes, and validations from natural language descriptions
• Producing idiomatic Ruby code with proper use of blocks, procs, and enumerable methods

Tips for AI-Assisted Ruby Development

• AI tools understand Rails conventions deeply - use standard naming and it will generate correct associations
• Leverage AI for generating RSpec tests with proper describe/context/it blocks
• AI handles ActiveRecord queries, scopes, and validations extremely well
• Use AI to generate Ruby metaprogramming patterns like concern modules and DSLs
• AI excels at generating Rake tasks and migration files

Prompting Tips for Ruby

Where AI Struggles with Ruby

• AI tools struggle with Ruby's heavy metaprogramming patterns (method_missing, define_method, class_eval), generating code that is syntactically correct but misses the meta-level intent
• AI has significantly less training data for non-Rails Ruby projects, resulting in lower-quality output for Sinatra, Hanami, dry-rb, or standalone Ruby applications
• Ruby's lack of static types means AI tools cannot verify generated code against contracts, leading to runtime NoMethodError and TypeError that only surface during execution
• AI often generates Rails code targeting outdated versions (Rails 5/6 patterns) instead of modern Rails 7+ conventions like Hotwire, Turbo, and importmaps

Rails service object with ActiveRecord query

A service object encapsulating business logic with an ActiveRecord query scope, demonstrating Rails patterns that AI tools generate with high accuracy.

class Orders::OverdueNotifier
  Result = Data.define(:notified_count, :errors)

  def initialize(mailer: OrderMailer, clock: Time)
    @mailer = mailer
    @clock = clock
  end

  def call
    errors = []
    overdue = Order.where(status: :pending)
                   .where("created_at < ?", 7.days.ago(@clock.current))
                   .includes(:customer)

    overdue.find_each do |order|
      @mailer.overdue_reminder(order).deliver_later
    rescue StandardError => e
      errors << { order_id: order.id, error: e.message }
    end

    Result.new(
      notified_count: overdue.count - errors.size,
      errors: errors
    )
  end
end

Common Use Cases

• Web applications with Ruby on Rails
• API backends and microservices
• DevOps scripting and automation
• Rapid prototyping and MVPs

Popular Ruby Libraries AI Handles Well

Best Practices

Follow Rails conventions strictly - AI tools rely on them for accurate code generation. Use yard documentation for better AI context. AI-generated Ruby tends to be idiomatic when the surrounding code follows Ruby community standards. Use Sorbet type annotations if available for improved AI assistance.

Recommended Tools for Ruby

The following AI coding tools offer the best support for Ruby development:

• Cursor - AI-first code editor built as a fork of VS Code with deep AI integration for code generation, editing, and chat.
• GitHub Copilot - AI pair programmer by GitHub and Microsoft that provides code suggestions, chat, and autonomous coding agents directly in your editor.
• Claude Code - Anthropic's agentic CLI coding tool that operates directly in your terminal, capable of editing files, running commands, and managing entire coding workflows.
• Cody - AI coding assistant by Sourcegraph that leverages deep codebase understanding and code search to provide context-aware assistance.

FAQ