dspy - Build and optimize declarative LM programs
Defines modular language-model programs, generates answers, retrieves context, and automatically optimizes prompts with data-driven methods.
Tags
Updated: 2026-09-29Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Define task signatures
- Compose modular pipelines
- Generate reasoning steps
- Retrieve relevant passages
- Use tools with agents
- Optimize prompts automatically
- Export fine-tuning data
Inputs
- Task signatures
- Training examples
- Evaluation metrics
- LM provider configuration
- Retrieval collections
- Tool functions
Outputs
- Generated answers and rationales
- Retrieved context passages
- Module predictions
- Optimized prompts and modules
- Fine-tuning datasets
Requirements
- DSPy package
- OpenAI or Anthropic packages
- Language model access
- Supported Linux, macOS, or Windows environment