DSPy – Program, don't prompt, your LLMs

DSPy 3.4.0 has been released with PythonInterpreter improvements, faster GEPA optimization, and MCP v2 compatibility, according to the Stanford NLP project's release notes. The Python framework, which has 38,000 GitHub stars, 461 contributors, and more than 5.2 million monthly downloads, lets developers define tasks as typed signatures and compile them against a metric; the release notes cite a GEPA example improving retrieval-augmented generation from 0.41 F1 to 0.63 F1.

DSPy 3.4.0 — PythonInterpreter improvements, faster GEPA, and MCP v2 compatibility · learn more → https://github.com/stanfordnlp/dspy/releases Program, don’t prompt, your LLMs. DSPy is a Python framework for building AI systems. Express your tasks as structured signatures, not prompts, to produce maintainable, modular, and optimizable programs. extract events.py 12345 678910 lm = dspy.LM "openai/gpt-5.4-nano" class ExtractEvent dspy.Signature : """Extract event details from an email.""" email: str = dspy.InputField event name: str = dspy.OutputField date: str = dspy.OutputField extract = dspy.Predict ExtractEvent extract email=inbox message output Prediction event name="Team Offsite", date="Thursday, June 5" 5.2M+ monthly downloads 461+ contributors 38k github stars in production at Compose programs with reusable primitives. Signatures Declare your task. Define your task as typed inputs and outputs instead of managing messy prompts. Portable, maintainable, and easy to iterate on. class Triage dspy.Signature : """Route a support ticket.""" ticket: str = dspy.InputField urgency: Literal "low", "high" = dspy.OutputField team: str = dspy.OutputField Modules Same interface, different strategy. Modules control how your signature executes. Reason, run ensembles, use tools, add a REPL, and more without rewriting your task. Direct completion classify = dspy.Predict Triage Add step-by-step reasoning classify = dspy.ChainOfThought Triage Add tools and a reasoning loop classify = dspy.ReAct Triage, tools= search Optimizers Compile your program against a metric. Give DSPy examples and a scoring function. It tunes your prompts automatically until quality converges. tp = dspy.GEPA metric=semantic f1, auto="medium" opt = tp.compile rag, trainset Before: 0.41 F1 After: 0.63 F1 opt.save "rag.v2.json" Define a task. Grow it into a system. class Extract dspy.Signature : """Extract contact info.""" message: str = dspy.InputField name: str = dspy.OutputField email: Optional str = dspy.OutputField intent: Literal "meeting", "intro", "follow-up" = dspy.OutputField extract = dspy.Predict Extract extract message="I'm Sarah" " sarah@acme.co . Meet Thursday?" outputstdout Prediction name="Sarah", email="sarah@acme.co", intent="meeting" def search query: str - list str : """Search a knowledge base.""" return kb.query query, k=3 def calc expr: str - float: """Evaluate a math expression.""" return dspy.PythonInterpreter {} .execute expr agent = dspy.ReAct "question - answer", tools= search, calc agent question="GDP per capita of France?" outputstdout thought 1: I need France's GDP and population. action 1: search "France GDP" → ... thought 2: Now divide GDP by population. action 2: calc "3.13e12 / 68e6" → 46029.4 Prediction answer="$46,029" class FactCheck dspy.Module : def init self : self.find = dspy.ChainOfThought "article - claims: list str " self.verify = dspy.ChainOfThought "claim, source - verdict" def forward self, article : found = self.find article=article return self.verify claim=c, source=article for c in found.claims outputstdout FactCheck article=news article Prediction verdict="supported" , Prediction verdict="unsupported" , Prediction verdict="supported" class AnalyzeChart dspy.Signature : """Describe the trend and key data points in a chart.""" chart: dspy.Image = dspy.InputField title: str = dspy.OutputField trend: str = dspy.OutputField data points: list dict = dspy.OutputField analyze = dspy.Predict AnalyzeChart analyze chart=dspy.Image "quarterly revenue.png" outputstdout Prediction title="Quarterly Revenue 2024 ", trend="Steady growth, Q3 dip, strong Q4 recovery", data points= {"q": "Q1", "rev": "$4.2M"}, ... optimizer = dspy.GEPA metric=accuracy, auto="medium" optimized = optimizer.compile extract, trainset=labeled emails optimized.save "extract v2.json" outputstdout Baseline 62% gpt-5.4-mini, zero-shot Optimized 89% gpt-5.4-mini + GEPA compile Cost $2.18 · 200 examples Saved to → extract v2.json Built in the open, since Dec 2022. DSPy started at Stanford NLP and grew into a research community. New optimizers and module types land here first — then show up in production systems at companies you’ve heard of. Dec 2025 Recursive Language Models Jul 2025 GEPA: Reflective Prompt Evolution Jul 2024 BetterTogether: Fine-Tuning + Prompt Opt. Jun 2024 MIPROv2: Optimizing Instructions & Demos Feb 2024 STORM: Writing Wikipedia-like Articles Oct 2023 DSPy: Compiling Declarative LM Calls Dec 2022 Demonstrate-Search-Predict DSPy in production Metadata extraction across all shops; ~550× cost reduction Optimized Dash relevance judge for ranking and evaluation Prompt migration from larger to smaller models on Amazon Nova Multiple chatbot use cases on Databricks Code repair pipeline using code LLMs to synthesize diffs LM judges, RAG, classification, and customer solutions Evolutionary self-improvement for the Hermes agent See all companies using DSPy in production