Getting Started with RubyLLM

This guide will help you get up and running with RubyLLM, showing you the basics of chatting with AI models, generating images, and creating embeddings.

Prerequisites

Before starting, make sure you have:

  1. Installed the RubyLLM gem (see the Installation guide)
  2. At least one API key from a supported provider (OpenAI, Anthropic, Google, or DeepSeek)

Basic Configuration

Let’s start by setting up RubyLLM with your API keys:

require 'ruby_llm'

RubyLLM.configure do |config|
  # Add the API keys you have available
  config.openai_api_key = ENV['OPENAI_API_KEY']
  config.anthropic_api_key = ENV['ANTHROPIC_API_KEY']
  config.gemini_api_key = ENV['GEMINI_API_KEY']
  config.deepseek_api_key = ENV['DEEPSEEK_API_KEY']
end

Your First Chat

Let’s start with a simple chat interaction:

# Create a chat (uses the default model)
chat = RubyLLM.chat

# Ask a question
response = chat.ask "What's the capital of France?"
puts response.content
# => "The capital of France is Paris."

# Continue the conversation
response = chat.ask "What's the population of that city?"
puts response.content
# => "Paris has a population of approximately 2.1 million people..."

Using a Specific Model

You can specify which model you want to use:

# Use Claude
claude_chat = RubyLLM.chat(model: 'claude-3-5-sonnet-20241022')
claude_chat.ask "Tell me about Ruby programming language"

# Use Gemini
gemini_chat = RubyLLM.chat(model: 'gemini-2.0-flash')
gemini_chat.ask "What are the best Ruby gems for machine learning?"

Exploring Available Models

RubyLLM gives you access to models from multiple providers. You can see what’s available:

# List all models
all_models = RubyLLM.models.all
puts "Total models: #{all_models.count}"

# List chat models
chat_models = RubyLLM.models.chat_models
puts "Chat models:"
chat_models.each do |model|
  puts "- #{model.id} (#{model.provider})"
end

# List embedding models
RubyLLM.models.embedding_models.each do |model|
  puts "- #{model.id} (#{model.provider})"
end

# Find info about a specific model
gpt = RubyLLM.models.find('gpt-4o-mini')
puts "Context window: #{gpt.context_window}"
puts "Max tokens: #{gpt.max_tokens}"
puts "Pricing: $#{gpt.input_price_per_million} per million input tokens"

Generating Images

RubyLLM makes it easy to generate images with DALL-E:

# Generate an image
image = RubyLLM.paint("a sunset over mountains")

# The URL where you can view/download the image
puts image.url

# How the model interpreted your prompt
puts image.revised_prompt

# Generate a larger image
large_image = RubyLLM.paint(
  "a cyberpunk city at night with neon lights",
  size: "1792x1024"
)

Creating Embeddings

Embeddings are vector representations of text that can be used for semantic search, classification, and more:

# Create an embedding for a single text
embedding = RubyLLM.embed("Ruby is a programmer's best friend")

# The vector representation
vector = embedding.vectors
puts "Vector dimension: #{vector.length}"

# Create embeddings for multiple texts
texts = ["Ruby", "Python", "JavaScript"]
embeddings = RubyLLM.embed(texts)

# Each text gets its own vector
puts "Number of vectors: #{embeddings.vectors.length}"

Working with Conversations

Here’s how to have a multi-turn conversation:

chat = RubyLLM.chat

# First message
chat.ask "What are the benefits of Ruby on Rails?"

# Follow-up questions
chat.ask "How does that compare to Django?"
chat.ask "Which one would you recommend for a new web project?"

# You can check all messages in the conversation
chat.messages.each do |message|
  puts "#{message.role}: #{message.content[0..100]}..."
end

What’s Next?

Now that you’ve got the basics down, you’re ready to explore more advanced features: