DSPy
Framework for programming—not prompting—language models
Founded
2022
Starting Price
$0
Category
Developer Tools
Last updated March 17, 2026
About DSPy
DSPy is an open-source framework from Stanford NLP that replaces manual prompt engineering with programmatic optimization. Define input/output signatures and let DSPy's optimizers automatically find the best prompts, few-shot examples, and fine-tuning strategies for your LLM pipeline.
Pros & Cons
Pros
- Eliminates manual prompt engineering — optimizers find better prompts automatically
- Programs are portable across LLM providers without rewriting prompts
- Stanford NLP backing ensures rigorous research foundation
- Reproducible results through programmatic optimization vs. ad-hoc prompting
- Completely free and open-source with MIT license
Key Features
Declarative Signatures
Define what your LLM should do with typed input/output specifications instead of hand-crafted prompts
Automatic Prompt Optimization
Optimizers like MIPROv2 automatically generate and test instructions and few-shot examples
GEPA Optimizer
Reflects on program trajectories to identify failures and propose prompt improvements
SIMBA Optimizer
Uses stochastic mini-batch sampling to find challenging examples and generate self-improvement rules
Compilation
Compile high-level code into optimized prompts or weight updates aligned with your metrics
Module Composition
Combine retrieval, generation, classification, and reasoning modules into complex pipelines
Multi-LLM Support
Works with OpenAI, Anthropic, Google, and local models through unified interface
Pricing
Open Source
$0
- Full framework
- All optimizers
- Community support
- MIT license
- Self-hosted
Best For
RAG Pipeline Optimization
Build and optimize retrieval-augmented generation systems with automatic prompt tuning for better accuracy
Classification Systems
Create LLM-based classifiers that automatically optimize prompts to maximize classification accuracy
Multi-Hop Reasoning
Build complex reasoning chains where DSPy optimizes each step's prompts for the overall pipeline metric
Research & Experimentation
Systematically explore prompt strategies and model configurations with reproducible optimization
Similar Tools
Featured In
#5
7 Best LangChain Alternatives for Building AI Applications (2026)
Best LangChain alternative for prompt optimization — the only framework that replaces manual prompt engineering with systematic, reproducible optimization
#1
7 Best AI Prompt Engineering & LLM Tools for Builders (2026)
Best for AI engineers building production LLM pipelines who want algorithmic prompt optimization — delivers measurable quality gains over manual engineering, but requires technical investment to learn.
