dspy

Define a metric function and training examples, then use a teleprompter like BootstrapFewShot or MIPRO to compile your module. The optimizer generates few-shot demonstrations and improved instructions from your data, producing an optimized module without manual prompt editing.

Predict makes a direct LM call for simple tasks, while ChainOfThought generates reasoning steps before the answer. ChainOfThought is roughly twice as slow but significantly more accurate on math and logical reasoning tasks.

DSPy optimizes prompts automatically from data and offers typed signatures with high modularity, while LangChain relies on manual prompts and prebuilt chains. Choose DSPy when you have training data and need systematic improvement; choose LangChain for quick prototypes with existing integrations.

Yes, DSPy supports local models through dspy.OllamaLocal by pointing to a local base URL such as http://localhost:11434. It also supports OpenAI, Anthropic Claude, and other providers, and you can mix models for different pipeline stages.

BootstrapFewShot works with 10-50 examples, MIPRO performs best with 50-200 examples plus a validation set, and BootstrapFinetune recommends 100 or more examples. Insufficient data is a common cause of poor optimization results.

Avoid DSPy for quick one-off prototypes where manual prompting is faster, or when you lack any training data for optimization. It also has a medium-to-high learning curve compared to direct prompting.