What DSPy does
DSPy is a Python framework for building AI systems. Developers express tasks as structured signatures with typed inputs and outputs instead of hand-written prompts. Named features include Signatures, Modules such as Predict, ChainOfThought, and ReAct, which can use tools, and Optimizers such as GEPA, which tune programs against a metric using examples. Programs can be saved and loaded, and images can serve as input fields for multimodal tasks. The page describes the goal as maintainable, modular, and optimizable programs but does not name a target audience. It requires Python 3.10 or later, installs with pip, and is released under the MIT license. The page states no pricing model. The project began at Stanford NLP.
How DSPy compares with its closest alternatives
3 of the 5 closest alternatives have a free plan, as DSPy does. All 5 offer an API, as DSPy does. DSPy offers a desktop app. Of the 3 alternatives with platforms on record, 1 offers a desktop app. DSPy was founded in 2023; the alternatives were founded between 2016 and 2023, 3 of them earlier. DSPy is based in Stanford, United States; 4 of its 5 alternatives are in the United States too, and the others in Japan.