dspy - Declarative LM Programs and Prompt Optimization
Build declarative language-model programs, optimize prompts with data, and create RAG systems, agents, or classifiers.
Tags
Updated: 2026-09-30Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Define task signatures
- Compose reusable modules
- Generate reasoning steps
- Optimize prompts with data
- Build RAG pipelines
- Configure language models
- Use tools with agents
- Export fine-tuning data
Inputs
- Task signatures and prompts
- Questions and context
- Training examples and metrics
- Retrieval collections
- LM provider configuration
- Tool functions
Outputs
- Predictions and answers
- Reasoning traces
- Optimized modules
- Fine-tuning datasets
Requirements
- DSPy package
- Supported LM provider package
- Language model access or Ollama
- Provider API credentials when required
- Linux, macOS, or Windows