MathReader
User guide使用指南
Read mathematical PDFs, find results and equations, extract LaTeX, and follow references between papers.
了解如何阅读数学 PDF、查找定理与公式、提取 LaTeX,以及关联参考文献。
1. Getting started1. 开始阅读
- Choose Open PDF on the welcome screen, or press O (or Ctrl/Cmd + O) when you are not typing.
- You can also drag a PDF into MathReader on desktop.
- To switch papers, press O again. On a touch device, choose History → Local library → Open PDF from the left panel.
- Previously cached papers appear in History and on the welcome screen when available.
- 在欢迎页点击打开 PDF,也可在未输入文字时按 O 或 Ctrl/Cmd + O。
- 在电脑上,也可以将 PDF 拖入 MathReader。
- 阅读中可再次按 O 切换论文;触屏设备可在左侧的历史 → 本地资料库 → 打开 PDF中选择文件。
- 已缓存的论文可在历史以及欢迎页的最近阅读中重新打开。
Use ? in the header to open this guide.
点击顶栏的 ? 可打开本指南。
2. Reader controls2. 阅读界面
Use the toolbar to change pages, enter a page number, zoom in or out, or switch between Fit width and Fit page. Arrow keys and Page Up/Down also navigate the PDF when you are not typing.
可用工具栏翻页、输入页码、缩放,或切换适合宽度与适合整页。未输入文字时,也可使用方向键和 Page Up/Down 浏览。
Use the left panel for Contents, Search, Bookmarks and History. The right panel provides notes, formula extraction and display settings. Collapse either panel whenever you need more reading space.
左侧栏提供目录、搜索、书签和历史;右侧栏提供笔记、公式提取与显示设置。需要更宽的阅读区域时,可收起任一侧栏。
3. Mathematical indexes3. 数学索引
Press C to open Contents. Browse Structure for sections, Results for theorems and proofs, Equations for numbered formulas, and Bibliography for cited works. Select an entry to jump to it. Your highlights and notes appear under Annotations, which opens automatically when you add one.
按 C 打开目录。结构用于查找章节,结果用于查找定理和证明,公式列出检测到的公式编号,参考文献列出引用文献。点击条目即可跳转。高亮和笔记会显示在批注中;新增批注时,该分组会自动展开。
Indexes are generated automatically and may miss entries, especially in scanned or low-quality PDFs. Check the original page when accuracy matters.
索引由程序自动生成;扫描件或画质较差的 PDF 可能存在遗漏。重要内容请以原文为准。
4. Link related papers4. 关联论文
You can connect a reference in one paper to a PDF already in your local library. You choose the matching file, so links take you to the paper you intended.
你可以把论文中的参考文献关联到本地资料库中已打开过的 PDF。由你确认目标文件,避免误跳转。
- Open both PDFs in MathReader once, then return to the paper containing the reference.
- Find the reference under Contents → Bibliography.
- Right-click the reference row and choose Link PDF. On touch devices, long-press the row.
- Choose the corresponding PDF from your library and confirm Link.
- To follow the reference later, right-click the row and choose Open PDF.
- Use ↩ Paper to return to where you were reading.
- 先分别在 MathReader 中打开两篇 PDF,再回到含有引用的论文。
- 在目录 → 参考文献中找到对应条目。
- 右键单击参考文献条目,选择关联 PDF;触屏设备可长按。
- 从本地资料库中选择对应 PDF,并确认关联。
- 以后可右键单击该条目,选择打开 PDF跳转。
- 点击↩ 原文即可回到跳转前的阅读位置。
Existing DOI and arXiv links continue to open their usual destinations. A local PDF link opens the selected paper at the document level.
原有 DOI 和 arXiv 链接仍可正常使用。本地关联会打开指定论文,不会自动定位其中的具体定理。
To change or remove a link, right-click the reference and choose Relink PDF or Unlink. Links are saved locally; the return trail resets when the reader reloads.
要修改或删除关联,可右键选择重新关联 PDF或取消关联。关联保存在本地;重新加载页面后,跨论文返回记录会清空。
5. Extract a formula as LaTeX5. 将公式提取为 LaTeX
- Open a PDF, press F to enable formula selection, and drag a rectangle around a formula.
- MathReader recognizes the selected image and returns editable LaTeX. The first extraction may take longer while the OCR model loads. For PDFs with selectable text, you can also compare the PDF vector result.
- Inspect the image crop, rendered math and editable LaTeX in the Formula OCR workspace. Correct incorrect symbols before copying or saving an equation.
- Press F again or Esc to exit selection mode.
- 打开 PDF,按 F 开启公式选取,在公式周围拖出矩形框。
- MathReader 会识别选中的公式并生成可编辑的 LaTeX。首次使用可能需要加载 OCR 模型,耗时稍长。有可选中文本的 PDF 也可用PDF 矢量结果作对照。
- 在右侧公式 OCR 中核对裁图、渲染结果和可编辑的 LaTeX;复制或收藏前应修正错误符号。
- 再按 F 或 Esc 退出选取模式。
Always check the symbols against the original formula, even if the LaTeX renders correctly. If recognition fails, try another engine in Settings → Formula OCR or use Reset OCR.
即使 LaTeX 能正常渲染,也请核对原公式的符号。识别失败时,可在设置 → 公式 OCR中切换引擎,或使用重置 OCR。
6. Scanned PDFs and document OCR6. 扫描版 PDF 与全文 OCR
When a scanned PDF has no selectable text, MathReader may offer Run OCR or Open without OCR. Full-document OCR takes extra time but can enable text-dependent Contents and search. Opening without it keeps the page readable but may leave some indexes unavailable. You can later run OCR from the Contents panel when prompted.
如果扫描版 PDF 无法选中文字,MathReader 可能提示运行 OCR或不运行 OCR 直接打开。全文 OCR 耗时较长,但可为目录与搜索提供文本。跳过 OCR 仍能看页面,但部分索引可能不可用;之后可根据目录面板提示再运行 OCR。
For a single formula, use formula selection instead of processing the whole document. Handwritten mathematics may require manual correction.
只需提取单个公式时,可使用公式选取,无需进行全文 OCR。手写数学内容通常需要人工核对。
7. Search and command palette7. 搜索和命令面板
Use Ctrl/Cmd + F to search the current PDF. Press Ctrl/Cmd + K for the command palette; type a command such as Open PDF, Bibliography or Rebuild mathematical structure index. MathReader also exposes known mathematical entries through the palette.
用 Ctrl/Cmd + F 搜索当前 PDF。按 Ctrl/Cmd + K 打开命令面板,可输入打开 PDF、参考文献或重新生成数学结构索引等命令。命令面板也可检索已识别的数学条目。
For image-only PDFs, run document OCR before searching text. Formula search is limited by what MathReader can recognize.
纯图片 PDF 需要先运行全文 OCR 才能搜索文字。公式搜索范围取决于识别结果。
8. Bookmarks, notes and pinned references8. 书签、批注与固定参考页
Select text to reveal actions such as Copy text, Highlight, Margin note, Page note and Glossary. Use N for the current page note, S for Bookmarks, and P to pin the current PDF page alongside your reading pane. A pinned page stays available for reference while you read elsewhere in the same PDF.
选中文字后可使用复制文本、高亮、页边注、页注、符号表等操作。按 N 打开当前页页注,按 S 打开书签,按 P 将当前 PDF 页固定在阅读区旁。固定页面可在阅读同一 PDF 的其他位置时作为参考。
Use Alt + ← to go back in the within-paper reading trail. Use ↩ Paper to return after opening a linked PDF.
按 Alt + ← 可返回文档内的阅读轨迹上一位置;通过参考文献打开其他 PDF 后,可用↩ 原文返回。
9. Local library, exports and privacy9. 本地资料库、导出与隐私
Open History for cached local PDFs and Open PDF on touch devices. MathReader keeps paper bytes and reading state in browser storage when caching is available. Export saves your reading data, not the original PDFs. Linked-reference mappings are not included.
在历史中查看已缓存的本地 PDF;触屏设备也可从这里选择打开 PDF。在缓存可用时,MathReader 将 PDF 和阅读状态保存在浏览器存储中。导出可保存阅读数据,但不包含原始 PDF 文件或参考文献的本地关联。
Clearing site storage or using a different browser/device can remove the cached papers and links. Keep original PDFs and independent backups. OCR assets may need to be downloaded; your PDF/formula crops are not uploaded to the MathReader website.
清理站点存储或换浏览器/设备可能导致缓存论文与关联消失。请保留原始 PDF 与独立备份。OCR 模型资源可能需要下载;你的 PDF 和公式裁图不会上传到 MathReader 网站。
10. Display and mobile layout10. 显示设置与移动端
Press D for display settings: continuous/single-page reading, page tone, brightness, contrast, spacing and shadows. The header switches language and light/dark themes. On narrow screens, the reader becomes full width: use ☰ Contents and ☷ Workspace to open panels and the visible Close × controls to dismiss them.
按 D 进入显示设置,可调整连续/单页阅读、纸张色调、亮度、对比度、间距和阴影。顶栏可切换语言与明暗主题。窄屏时阅读区铺满宽度;使用☰ 目录、☷ 工作区展开面板,再点击关闭 ×收起。
11. Keyboard shortcuts11. 快捷键速查
Shortcuts work while you are reading, not typing in a field. Some may be handled by your browser.
在阅读区使用快捷键;输入文字时不生效。部分组合键可能由浏览器接管。
Document文档
- Open PDF打开 PDF
- O
- Search PDF搜索 PDF
- Ctrl/Cmd + F
- Previous / next page上一页 / 下一页
- ← / →
- Scroll滚动页面
- ↑ / ↓
Reading & notes阅读与批注
- Select formula选取公式
- F
- Bookmarks书签
- S
- Page note页注
- N
- Pin page固定页面
- P
- Back to previous paper返回上一论文
- Alt + ←
- Reference menu参考文献菜单
- Shift + F10
Workspace工作区
- Contents目录
- C
- History / library历史 / 资料库
- H
- Left sidebar左侧栏
- L
- Right sidebar右侧栏
- R
- Both sidebars (desktop)两侧栏(桌面)
- B
- Commands命令面板
- Ctrl/Cmd + K
- Display settings显示设置
- D
12. Troubleshooting12. 常见问题
How do I open another PDF?
Use O, Ctrl/Cmd + K → Open PDF, drag-and-drop, or History → Local library → Open PDF (touch-friendly).
How do I link a reference to a local PDF?
Right-click the reference row and choose Link PDF; on a phone, long-press it. After linking, use the same menu to open the PDF.
My linked paper is not listed or cannot open.
Open the exact target PDF once in MathReader and ensure its bytes are cached. If it was forgotten or browser storage was cleared, open it again and relink manually.
Why are some sections or equations missing?
Check whether the PDF has a usable text layer. Use document OCR for scanned material where offered; mathematical detection is intentionally conservative and imperfect.
What if formula recognition fails?
Wait for the first image-model initialization, check Settings → Formula OCR → Troubleshooting → Reset OCR, or try a CPU engine. Vector extraction may be used for comparison when PDF text exists.
How can I avoid losing my reading data?
Browser-local data depends on that browser's storage. Maintain separate backups of original PDFs and exported reading data; Linked-reference mappings are not included in the export.
如何打开其他 PDF?
使用 O、Ctrl/Cmd + K → 打开 PDF、拖入文件,或进入 历史 → 本地资料库 → 打开 PDF(适合触屏)。
如何把参考文献关联到本地 PDF?
右键单击参考文献条目,选择关联 PDF;手机上可长按。完成关联后,可从同一菜单打开论文。
找不到关联目标,或目标无法打开。
先在 MathReader 中打开确切的目标 PDF,确认已缓存。如曾忘记该文件或清理站点存储,请重新打开并手动关联。
为什么部分章节或公式没有出现在索引中?
确认 PDF 是否存在可用文本层。扫描材料可按提示运行全文 OCR;数学索引采用保守检测,无法保证完全无误。
公式识别失败怎么办?
首次使用可能需要初始化模型;可在 设置 → 公式 OCR → 故障处理 中重置 OCR,或改用 CPU 引擎。有文本层的 PDF 可以用矢量提取作对照。
怎样避免丢失阅读数据?
本地数据依赖浏览器存储。请独立备份原始 PDF 和导出的阅读数据;参考文献关联目前不包含在导出内容中。