离线 · 无上传 · 浏览器本地完成Offline · no uploads · runs in the browser
PDF 转 JPEG——本地逐页栅格化，清晰度与质量可调PDF to JPEG locally — one page per image, density set
这条和 PDF → PNG 是同一套渲染，区别只在最后那一次编码：本项目给 JPEG 单独开了一条直接边，而不是"先出 PNG 再转 JPEG"。原因很实际——中间那次无损编码不改变任何像素，只会让每个页面多走一遍编码、并让 ZIP 里的扩展名变成 .png。栅格化是像素这条路唯一的真瓶颈，所以能省的步骤就省掉。This shares its whole render with PDF to PNG and differs only in the final encode: the extension gives JPEG its own direct edge rather than going PDF → PNG → JPEG. The reason is mundane — the lossless middle step changes no pixels, costs one extra encode per page, and would leave a ZIP full of .png files. Rasterising is the only real cost on a pixel route, so the wasted step is not taken.
这条链路怎么走How the route runs
PDFPDF JPEGJPEG
共 1 步，由扩展的路径搜索自动选出，中间产物不落盘。1 step, chosen by the extension's own path search; intermediate results never touch the disk.
会保留什么What carries over
- 每一页独立成一张图，按 PDF 里的页码命名（`page-3.jpg` 就是第 3 页，只转第 3 页时也一样）Each page becomes its own image, named by its page number in the document — `page-3.jpg` is page 3, and stays page 3 when you convert only that page
- 渲染密度四档可选：96 / 144 / 200 / 300 DPI，默认 144；数值直接换算成 pdf.js 的页面缩放（DPI ÷ 72）Render density comes as 96, 144, 200 or 300 DPI with 144 as the default, converted straight into pdf.js page scale (DPI ÷ 72)
- 页面自身的矢量图形、文字的字形、颜色与渐变都按矢量渲染路径栅格化，不是截屏那种二次放大Vector shapes, letterforms and colour or gradients on the page are rasterised along the vector render path, not magnified from a screenshot
- 质量档 40%–90% 与目标体积只在链路末尾这次编码生效，中间步骤不吃这两个参数Quality steps of 40%–90% and a target size act only at this final encode — intermediate steps ignore both
- 超过浏览器画布上限（8192 px 一边）的页面自动按缩放上限收敛，是缩到能渲染，不是裁掉A page that would exceed the browser canvas limit (8192 px on an edge) is scaled down to fit rather than cropped
如实说明的限制What it does not do
- 产物是像素：文字不再是可选中、可复制的字符，也不能再被搜索；想保留文字层就转 HTML 或纯文本The output is pixels: text is no longer selectable, copyable or searchable. Keep the text layer by going to HTML or plain text instead
- 透明不保留：渲染前每页先铺一层白，而这条路的三个图片终点（PNG / JPEG / WebP）都是这样画的，所以从 PDF 出来的图一律没有透明区域Transparency does not survive: each page is painted white before it renders, and all three image targets (PNG, JPEG, WebP) are drawn that way, so a picture out of a PDF has no transparent areas
- 有损编码：默认质量下小字号与细线会先变糊，需要清晰锐利请选 PNG 或提高 DPIIt is a lossy encode: at the default quality small type and hairlines soften first. Choose PNG, or raise the density, when sharpness is the requirement
- PDF 内嵌的表单控件、批注、书签与附件不进图像；页面里"看起来是图片"的部分会跟着渲染进去，但无法再按对象取出Form fields, annotations, bookmarks and embedded files are not part of the image; whatever looks like a picture on the page is rendered in, but cannot be lifted out as an object
- 扫描件在这里只是分辨率更高的图片，仍然提取不出文字——项目不含 OCRA scan comes back as a higher-resolution picture of a scan, still without extractable text — the extension ships no OCR
实操提示Working with it
- 只想转其中几页：在页面范围里填 1,3-5 这类写法。填了的范围在这份文档里一页都对不上时会直接报错，而不是"干脆全给你"To convert only some pages, type a range such as 1,3-5. A range that matches no page in this document fails outright instead of quietly handing you the whole file
- 单页 PDF 直接得到一张 .jpg；两页以上才打包成 ZIPA one-page PDF gives you one .jpg directly; two or more pages arrive inside a ZIP
- 一批最多 200 个文件、单个文件 100 MB 上限；页数多时可以先用页面范围试一页，确认清晰度再全量跑Up to 200 files per batch with a 100 MB per-file ceiling; on a long document, render a single page first to check sharpness before running the lot
常见问题Questions people actually ask
DPI 该设多少？What density should I pick?
给人看、发群里：默认 144 就够，它与这个转换在有了可调密度之前所用的缩放完全一致。要打印、要放大看清小字：选最高的 300。档位越高像素越多、ZIP 越大，而页面内容不会因此多出任何细节——它只是把同一套矢量画得更密。For reading and messaging: the default 144, which is exactly the scale this converter used before the option existed. For print or zooming into small type: the top step, 300. More density means more pixels and a bigger ZIP, not more detail in the page — it draws the same vectors at a finer grid.
为什么转出来全是白底？Why is the background white?
因为渲染前每页都会先铺一层白。这是这条转换自己的决定，不是渲染出错，而且 PNG 与 WebP 同样先铺白——换终点换不回透明，从 PDF 出来的三种图片都没有透明区域。Because each page is painted white before it renders. That is this converter’s own decision, not a rendering fault, and PNG and WebP are painted white as well — changing the target does not win alpha back, and none of the three image outputs from a PDF has transparent areas.
100 页的 PDF 一次跑完会怎样？What happens if I run a 100-page PDF?
会得到一个含 100 张图的 ZIP，而且这是这条路径最重的用法——每页一张画布、逐页编码。所以这一路支持中途取消：取消后已完成的部分不会伪装成"成功"，界面会明确标成被中断。真嫌慢就用页面范围只跑要的页。You get a ZIP holding 100 images, which is the heaviest shape this route has: one canvas per page, encoded page by page. That is why it answers cancellation while it works, and a cancelled run is reported as interrupted rather than dressed up as success. If the wait is the problem, narrow the page range to the pages you need.
不需要账号，也不需要联网：装好扩展后点图标，在新标签页里打开工作台，文件从磁盘直接选。No account and no network: install the extension, click its icon, and the workbench opens in a new tab reading straight off your disk.
从 Chrome 应用商店安装Install from the Chrome Web Store 安装步骤Install steps GitHub 源码Source on GitHub
要把它压到指定体积？Need it under a target size? The 图片压缩工具image compressor可以在浏览器里重编码 PNG / JPEG / WebP。 re-encodes PNG / JPEG / WebP in the browser.