Your First Chat

The heart of RubyLLM is RubyLLM.chat. Three lines of code is all you need:

chat = RubyLLM.chat
response = chat.ask("What is Ruby?")
puts response.content

RubyLLM picks a sensible default model, sends your prompt, and returns a response object.

The Response Object

Every call to chat.ask(...) returns a response with useful metadata:

Your Task

Open first_chat.rb and:

1. Create a chat with RubyLLM.chat
2. Ask a question with chat.ask("...")
3. Print the response content and metadata

chat = RubyLLM.chat
response = chat.ask("What makes Ruby special as a programming language? Answer in 2-3 sentences.")

puts response.content
puts
puts "Model:         #{response.model_id}"
puts "Input tokens:  #{response.input_tokens}"
puts "Output tokens: #{response.output_tokens}"

Run it:

$ ruby first_chat.rb

Choosing a Model

Pass a model ID to use a specific model:

chat = RubyLLM.chat(model: "claude-sonnet-4-6")  # Anthropic
chat = RubyLLM.chat(model: "gpt-4o")             # OpenAI
chat = RubyLLM.chat(model: "gemini-2.0-flash")   # Google

You can even switch models mid-conversation:

chat = RubyLLM.chat(model: "gpt-4o")
chat.ask("What is Ruby?")

chat.with_model("claude-sonnet-4-6")
chat.ask("Now explain it differently")

The conversation history carries over — the new model sees everything the previous one said.