{
  "schemaVersion": "1.0",
  "updates": [],
  "companyNews": [],
  "usageGuide": {
    "facts": [
      {
        "key": "limits",
        "label": "使用前看",
        "text": "代码为 Apache 2.0，权重为修改版 OpenRAIL-M。权重条款限制前一年收入或累计融资超过 500 万美元主体的非个人／非研究使用，并单独限制竞品用途；须核对原文或另取许可。可选外部 LLM 另有传输和费用条件。",
        "sourceUrls": [
          "https://github.com/datalab-to/marker/blob/master/README.md",
          "https://github.com/datalab-to/marker/blob/master/LICENSE",
          "https://github.com/datalab-to/marker/blob/master/MODEL_LICENSE"
        ]
      }
    ]
  },
  "id": "b929",
  "language": "zh",
  "name": "Marker",
  "originalName": "Marker",
  "library": "tools",
  "category": "模型开发",
  "categoryKey": "模型开发",
  "categoryKeys": [
    "模型开发"
  ],
  "categories": [
    "模型开发"
  ],
  "entryPoints": [],
  "summary": "把 PDF 与文档转成 Markdown、JSON 或内容块，可选模型增强表格和公式处理。",
  "introduction": "Marker 是 Datalab 维护的文档转换工具，可从 PDF、图像和 Office 等格式提取正文、表格、公式与图像。它提供 fast 和 balanced 路径，按设备和文档情况组合文本层、版式模型、OCR 与视觉模型，并可额外接入 LLM 改善跨页表格等结果。托管 Datalab 平台使用另一个模型 Chandra，不能把托管平台的效果或数据承诺直接当作 Marker 的表现。",
  "capabilities": [],
  "company": {
    "name": "Endless Labs, Inc. dba Datalab",
    "originalName": "Endless Labs, Inc. dba Datalab",
    "country": {
      "codes": [
        "US"
      ],
      "names": [
        "美国"
      ],
      "display": "美国",
      "basis": "company-location",
      "basisLabel": "公开公司地址"
    }
  },
  "officialUrl": "https://github.com/datalab-to/marker",
  "profileUpdatedOn": "2026-10-23",
  "sources": [
    {
      "kind": "product",
      "url": "https://github.com/datalab-to/marker"
    },
    {
      "kind": "documentation",
      "url": "https://github.com/datalab-to/marker/blob/master/README.md"
    }
  ],
  "news": [],
  "links": {
    "interactive": "products/marker--b929.html",
    "reading": "read/zh/products/marker--b929.html",
    "markdown": "read/zh/products/marker--b929.md",
    "json": "read/zh/products/marker--b929.json"
  }
}