dspy

Name: dspy
Availability: InStock
Author: rnben

Build declarative AI pipelines orchestrating LM tasks across providers.

1|Updated Apr 18, 2026

Search Tags:#optimization #ai #build #pipeline #declarative #modular #lm

Ready to run

npx skills add https://github.com/rnben/hermes-skills --skill dspy-rnben

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: dspy
Source: https://github.com/rnben/hermes-skills/tree/main/plugins/mlops-skills/skills/dspy
Command: npx skills add https://github.com/rnben/hermes-skills --skill dspy-rnben

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

DSPy enables developers to design and orchestrate end-to-end AI systems using a declarative programming model, replacing manual prompt engineering with structured pipelines.

Core Features & Use Cases

Structured Signatures to define inputs/outputs for AI tasks.
Modular Modules & Pipelines including Predict, ChainOfThought, ReAct, and ProgramOfThought for building complex workflows.
Optimizers & RAG Support to improve prompts and integrate retrieval-augmented generation across providers.
Use cases include building multi-stage Q&A, document QA, and agent-based systems across OpenAI, Anthropic, or local models.

Quick Start

Provide a task goal and DSPy will assemble a modular LM workflow.

Frequently Asked Questions about dspy

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build declarative AI pipelines for multi-stage Q&A and agent workflows?▼

You can build declarative AI pipelines by defining structured signatures for inputs and outputs, then assembling modular components like Predict, ChainOfThought, and ReAct to orchestrate complex multi-stage Q&A and agent workflows.

What is the best way to replace manual prompt engineering when building modular LM systems?▼

Replacing manual prompt engineering is best achieved using declarative programming models with structured signatures and optimizers, which automatically improve prompts and orchestrate tasks across diverse language model providers.

Can I use ChainOfThought and ReAct modules with local models or specific providers like OpenAI?▼

Yes, you can use modules like ChainOfThought and ReAct with local models, OpenAI, or Anthropic by applying provider configurations to orchestrate tasks across diverse deployment scenarios.

How do I set up retrieval-augmented generation and optimize prompts in a declarative LM pipeline?▼

To set up retrieval-augmented generation, you integrate RAG support within your declarative pipeline and apply optimizers to systematically improve prompts and enhance the accuracy of document QA tasks.

Does ProgramOfThought work for building complex agent-based systems across different providers?▼

ProgramOfThought works for building complex agent-based systems by allowing you to define structured task signatures and orchestrate modular workflows seamlessly across multiple language model providers.

When should I use declarative LM pipelines instead of manual prompt construction for AI tasks?▼

You should use declarative LM pipelines instead of manual prompt construction when building scalable, modular AI systems like multi-stage Q&A or agent workflows, replacing fragile manual prompts with structured, optimizable signatures.