> ## Documentation Index
> Fetch the complete documentation index at: https://docs.somark.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# SoMark 简介

> SoMark 产品能力与核心特性概览

# 欢迎使用 SoMark

SoMark准确、快速、经济的将PDF、PPT、图片等各场景下多种格式文档转换为机器可读的结构化结果，为LLM的训练及RAG等应用提供数据弹药。

<CardGroup cols={3}>
  <Card title="99% 识别准确率" icon="bullseye">
    基础文字识别准确率达到 99%，支持原文坐标回溯，精确定位每个元素。
  </Card>

  <Card title="100 页最快 5 秒" icon="bolt">
    高速解析，支持集群横向扩展，轻松应对大规模批量处理需求。
  </Card>

  <Card title="按需付费" icon="circle-dollar-to-slot">
    按量计费或一次性授权，私有化部署最低仅需一张 RTX 3090 显卡。
  </Card>

  <Card title="21 种文档组件" icon="layer-group">
    识别标题、表格、公式、图片、化学结构式、印章、二维码等 21 种元素类型。
  </Card>

  <Card title="多格式输出" icon="file-export">
    输出 Markdown、JSON，适配训练数据与 RAG 应用场景。
  </Card>

  <Card title="多场景适配" icon="files">
    支持论文、研报、白皮书、合同、教育题库、政府文件等主流文档场景。
  </Card>
</CardGroup>

## 支持的文件格式

`pdf` `png` `jpg` `jpeg` `bmp` `tiff` `jp2` `dib` `ppm` `pgm` `pbm` `gif` `heic` `heif` `webp` `xpm` `tga` `dds` `xbm` `doc` `docx` `ppt` `pptx` `xlsx` `xlsm` `xls`

## 支持的文档元素

支持识别以下 21 种文档元素：

| 类别      | 元素                                                            |
| ------- | ------------------------------------------------------------- |
| 文本结构    | 标题 `title`、文字段 `text`、页眉 `header`、页脚 `footer`、脚注 `footnote`   |
| 图片与表格   | 图片 `figure`、图例 `figure_caption`、表格 `table`、表例 `table_caption` |
| 专业内容    | 公式 `equation`、化学结构式 `cs`、化学方程式 `cs_equation`、代码段 `code`       |
| 导航与版面   | 侧边栏 `sider`、目录 `cate`、目录条目 `cate_item`                        |
| 教育与结构化项 | 选项 `choice`、填空空白 `blank`、参考文献 `reference`                     |
| 特殊元素    | 二维码 `qrcode`、印章 `stamp`                                       |

<div style={{ display: 'grid', gridTemplateColumns: 'repeat(7, minmax(0, 1fr))', gap: '12px' }}>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E6%A0%87%E9%A2%98.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=d65bd8ea3ce62a07f8cb0fa9902322c3" alt="标题" width="28" data-path="images/标题.png" /><div>标题</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E6%96%87%E5%AD%97%E6%AE%B5.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=1f73de67c708804e020b3faeb6aabf72" alt="文字段" width="28" data-path="images/文字段.png" /><div>文字段</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%9B%BE%E7%89%87.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=a9128b2e633916639c58a426b23fa567" alt="图片" width="28" data-path="images/图片.png" /><div>图片</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%9B%BE%E4%BE%8B.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=edc155b44b9eb820feaca2fb700cfa48" alt="图例" width="28" data-path="images/图例.png" /><div>图例</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E8%A1%A8%E6%A0%BC.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=6e27e319f25e0a2576c335d0f05d9431" alt="表格" width="28" data-path="images/表格.png" /><div>表格</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E8%A1%A8%E4%BE%8B.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=5608e354e6bc0d0bd5518df25f62ce93" alt="表例" width="28" data-path="images/表例.png" /><div>表例</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%85%AC%E5%BC%8F.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=d24dd877000fd17f084a05e4f4bae078" alt="公式" width="28" data-path="images/公式.png" /><div>公式</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E9%A1%B5%E7%9C%89.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=2527b3fc3f6c489301b71464d58d94eb" alt="页眉" width="28" data-path="images/页眉.png" /><div>页眉</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E9%A1%B5%E8%84%9A.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=9d2842e6f26775670fe4c0c5f72da276" alt="页脚" width="28" data-path="images/页脚.png" /><div>页脚</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E4%BE%A7%E8%BE%B9%E6%A0%8F.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=d480e9c435541f26808968c25089789f" alt="侧边栏" width="28" data-path="images/侧边栏.png" /><div>侧边栏</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E8%84%9A%E6%B3%A8.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=01aabef7cd12227e2cee878d0e92cd54" alt="脚注" width="28" data-path="images/脚注.png" /><div>脚注</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E7%9B%AE%E5%BD%95.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=2f25585c3c882364e442dd3cccfe8e34" alt="目录" width="28" data-path="images/目录.png" /><div>目录</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E7%9B%AE%E5%BD%95%E6%9D%A1%E7%9B%AE.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=62f6f7fdaa17f52ac37a66b5e412b95f" alt="目录条目" width="28" data-path="images/目录条目.png" /><div>目录条目</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E9%80%89%E9%A1%B9.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=2763e8192daf06c093475ce28584fda0" alt="选项" width="28" data-path="images/选项.png" /><div>选项</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E4%BB%A3%E7%A0%81%E6%AE%B5.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=e7518732919406799f1454c1e25f3ad3" alt="代码段" width="28" data-path="images/代码段.png" /><div>代码段</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%A1%AB%E7%A9%BA%E7%A9%BA%E7%99%BD.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=fcc0ff7bc607373d6b58383ea7df4b51" alt="填空空白" width="28" data-path="images/填空空白.png" /><div>填空空白</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%8F%82%E8%80%83%E6%96%87%E7%8C%AE.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=8485137a7dec6990950fc4cb44d11295" alt="参考文献" width="28" data-path="images/参考文献.png" /><div>参考文献</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E4%BA%8C%E7%BB%B4%E7%A0%81.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=e8aca4d398bdd944d8fcd65c072f7c0a" alt="二维码" width="28" data-path="images/二维码.png" /><div>二维码</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%8D%B0%E7%AB%A0.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=d1588e292e16b1ad8ec9da5691749caf" alt="印章" width="28" data-path="images/印章.png" /><div>印章</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%8C%96%E5%AD%A6%E7%BB%93%E6%9E%84%E5%BC%8F.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=de44671c6e0b2db85a647eab49a1d24e" alt="化学结构式" width="28" data-path="images/化学结构式.png" /><div>化学结构式</div></div>
  <div style={{ border: '1px solid var(--gray-200)', borderRadius: '12px', padding: '12px', textAlign: 'center', display: 'flex', flexDirection: 'column', alignItems: 'center' }}><img src="https://mintcdn.com/soulcode-9c5983db/Ej60brLPNq6nqtN4/images/%E5%8C%96%E5%AD%A6%E6%96%B9%E7%A8%8B%E5%BC%8F.png?fit=max&auto=format&n=Ej60brLPNq6nqtN4&q=85&s=3862b4b97ab9f03a6b7afefc2a565537" alt="化学方程式" width="28" data-path="images/化学方程式.png" /><div>化学方程式</div></div>
</div>

## 快速开始

请参阅 [入门总览](/documentation/get-started-overview) 开始使用；如果你想先看接口能力和限制，可直接跳到 [API 概览](/api-reference/index)，常见问题见 [FAQ](/qa)。
