D

DSPy

Programming framework for optimizing LLM prompts and weights

FrameworkOpen SourceEarly
Visit Website18K

About

DSPy by Stanford NLP provides a programming model for building LLM applications by replacing prompting with programming. It automatically optimizes prompts and weights through compilation, making LLM applications more systematic and reproducible.

Strengths

  • Novel programming paradigm
  • Automatic prompt optimization
  • Research-backed

Limitations

  • Steep learning curve
  • Different mental model
  • Smaller ecosystem

Use Cases

Prompt optimizationLLM pipelinesResearch

Integrations

OpenAIAnthropicLocal models