stanfordnlp/dspy

Health v2 · maintenance only

Adoption not part of the score

38406 stars · 3375 forks observed · 2026-09-29

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-28, confidence 0.97

DSPy is a Python library that lets you program language models using declarative signatures instead of hand‑written prompts. It provides modules for building RAG pipelines, agents, and optimizable LLM programs.

Use cases

• build a multi‑step RAG pipeline with retrieval and generation
• create an LLM‑driven chatbot that can call external tools
• optimize prompts and weights automatically for higher quality outputs
• prototype AI agents with reasoning and tool‑use without writing prompt strings
• compose modular AI systems that are easy to test and maintain

When to choose

• you need a maintainable, composable way to build LLM applications
• you want automatic prompt optimization and metric‑driven improvement
• your project involves multi‑stage pipelines, tool‑calling, or agents

When to avoid

• you only need a single simple prompt and prefer a lightweight API wrapper
• your environment does not support Python 3.10+ or the required dependencies
• you require deep integration with non‑LLM specific frameworks (e.g., pure web frameworks)

Facets

library · maturity active

llm-inference agent-framework rag artificial-intelligence large-language-models chatbots python windows declarative-llm-programming modular-ai-systems prompt-optimization tool-integration signature-based retrieval-augmented-generation linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="stanfordnlp/dspy")