DeepYardDeepYard

DSPy vs Microsoft Agent Framework

Side-by-side comparison with live GitHub signals. Last updated May 16, 2026.

D

DSPy

Framework for programming — not prompting — LLMs — 33K+ stars

OSSFree
34.5Ktoday393
M

Microsoft Agent Framework

Unified multi-agent framework merging AutoGen and Semantic Kernel

OSSFree
58.1K800.0K/w4w ago445
MetricDSPyMicrosoft Agent Framework
GitHub Stars34.5K58.1K
Contributors393445
Last CommitMay 15, 2026Apr 15, 2026
Open Issues474823
Licenseopen-sourceopen-source
Pricingopen-sourceopen-source
Free TierYesYes
Categoryframeworksframeworks
TrendingNoNo

Shared Tags

pythonopen-source

Only in DSPy

prompt-optimizationprogrammingresearchragstanford

Only in Microsoft Agent Framework

microsoftmulti-agententerprisedotnet

About DSPy

DSPy replaces prompt engineering with programming. Instead of writing prompts, you define modules with input/output signatures and DSPy automatically optimizes the prompts and weights for your pipeline. Supports chain-of-thought, retrieval-augmented generation, and multi-hop reasoning patterns. From Stanford NLP.

View full listing

About Microsoft Agent Framework

Microsoft's Agent Framework is the convergence of AutoGen and Semantic Kernel into a unified multi-agent platform. It provides a consistent API for building, orchestrating, and deploying agent teams across Python and .NET, with enterprise features like managed identity, Azure integration, and production-grade observability.

View full listing