class Riffer::Config

Configuration for the Riffer framework.

Provides configuration options for AI providers and other settings.

Riffer.config.openai.api_key = "sk-..."

Riffer.config.amazon_bedrock.region = "us-east-1"
Riffer.config.amazon_bedrock.api_token = "..."

Riffer.config.anthropic.api_key = "sk-ant-..."

Riffer.config.evals.judge_model = "anthropic/claude-sonnet-4-20250514"

Constants

AmazonBedrock
Anthropic
AzureOpenAI
Evals
Gemini
OpenAI

Attributes

amazon_bedrock [R]

Amazon Bedrock configuration (Struct with api_token and region).

anthropic [R]

Anthropic configuration (Struct with api_key).

azure_openai [R]

Azure OpenAI configuration (Struct with api_key and endpoint).

evals [R]

Evals configuration (Struct with judge_model).

gemini [R]

Google Gemini configuration (Struct with api_key, open_timeout, and read_timeout).

openai [R]

OpenAI configuration (Struct with api_key).

tool_runtime [R]

Global tool runtime configuration (experimental).

Accepts a Riffer::ToolRuntime subclass, a Riffer::ToolRuntime instance, or a Proc. Defaults to Riffer::ToolRuntime::Inline.new.

Public Class Methods

new ()

Source

# File lib/riffer/config.rb, line 64
def initialize
  @amazon_bedrock = AmazonBedrock.new
  @anthropic = Anthropic.new
  @azure_openai = AzureOpenAI.new
  @gemini = Gemini.new
  @openai = OpenAI.new
  @evals = Evals.new
  @tool_runtime = Riffer::ToolRuntime::Inline.new
end

Public Instance Methods

tool_runtime= (value)

Source

# File lib/riffer/config.rb, line 56
def tool_runtime=(value)
  valid = (value.is_a?(Class) && value < Riffer::ToolRuntime) || value.is_a?(Riffer::ToolRuntime) || value.is_a?(Proc)
  raise Riffer::ArgumentError, "tool_runtime must be a Riffer::ToolRuntime subclass, instance, or a Proc" unless valid
  @tool_runtime = value
end

Sets the global tool runtime.

Raises Riffer::ArgumentError if the value is not a valid runtime (ToolRuntime subclass, ToolRuntime instance, or Proc).