dspy-ruby

You can build type-safe LLM applications in Ruby by defining typed signatures using T::Struct and T::Enum, which ensures deterministic prompts and structured outputs. This approach provides modular composition of predictors and tools for reliable execution.

DSPy.rb is used for orchestrating autonomous tasks with agents in Ruby apps by providing signature and module tooling. It enables defining structured signatures, composing modular components, and configuring language model providers per tool.

Yes, you can configure different language models for individual tools through per-tool LM configuration. This allows you to assign specific providers to distinct modular components within your Ruby LLM application architecture.

DSPy.rb supports observability and optimization integrations including Langfuse, GEPA, and MIPROv2. These integrations enable monitoring and optimizing your type-safe Ruby LLM applications across modular components.

You define modular signatures and predictors by using DSPy.rb's DSLs for tools and modules alongside the typed type system. This enables modular composition of signatures, modules, and predictors for structured LLM outputs.

Typed signatures require using T::Struct and T::Enum for deterministic prompts and outputs, meaning your Ruby project must support this typed type system. This constraint ensures structured outputs but requires strict type definitions across your modular components.