{
  "schemaVersion": "1.0",
  "updates": [],
  "companyNews": [],
  "id": "b676",
  "language": "en",
  "name": "DSPy",
  "originalName": "DSPy",
  "library": "tools",
  "category": "Model development",
  "categoryKey": "模型开发",
  "categoryKeys": [
    "模型开发"
  ],
  "categories": [
    "Model development"
  ],
  "entryPoints": [],
  "summary": "Build composable Python modules for language-model tasks and optimize prompts or weights using examples and metrics.",
  "introduction": "DSPy expresses language-model programs with input/output signatures and composable modules for classification, RAG and agent loops. Developers supply examples and metrics, then choose optimizers for prompts, demonstrations or supported weight-training workflows. Iteration becomes an evaluable development process; improvements depend on data, objectives and the selected model.",
  "capabilities": [],
  "company": {
    "name": "Stanford NLP / DSPy contributors",
    "originalName": "Stanford NLP / DSPy contributors"
  },
  "officialUrl": "https://dspy.ai/current/",
  "profileUpdatedOn": "2026-10-22",
  "sources": [
    {
      "kind": "product",
      "url": "https://dspy.ai/current/"
    }
  ],
  "news": [],
  "links": {
    "interactive": "products/dspy--b676.html",
    "reading": "read/en/products/dspy--b676.html",
    "markdown": "read/en/products/dspy--b676.md",
    "json": "read/en/products/dspy--b676.json"
  }
}