DSPy
Stanford framework for programming (not prompting) LMs
Verdict
Game-changer for prompt optimisation. Define signatures and let DSPy optimise them with a dataset. Steep learning curve but worth it for production pipelines that need consistent quality.
Other Frameworks & SDKs
- GuidanceExperimental
Microsoft's constrained generation library for structured LLM output
- HaystackStable
Deepset open-source NLP framework for production LLM pipelines
- InstructorProduction
Structured outputs from LLMs using Pydantic schemas
- LMQLExperimental
Query language for LLMs — SQL-like syntax for constrained generation
- LangChainProduction
Composable LLM application framework — chains, agents, RAG
- LlamaIndexProduction
Data framework for LLM applications — indexing and retrieval

