打开 PDFnoted 网站，点击「OCR 扫描件识别」工具入口；拖放或点击上传你的扫描版PDF（支持A4常见分辨率，推荐300 DPI以上）。PDFnoted 自动检测是否为图像型PDF，若确认为扫描件，将启用OCR引擎。
在识别界面选择主要文字语言（如「简体中文+英文」双语模式），点击「开始识别」。PDFnoted 基于深度学习模型处理每页，平均3–8秒完成整份文档识别，结果实时高亮显示识别区域与置信度。
识别完成后，点击「导出为文本」获取纯TXT（保留段落结构），或点击「导出为可搜索PDF」生成带隐藏文字图层的新PDF——既保留原扫描样式，又支持全文搜索、文字选取与复制。所有文件均本地处理，不上传服务器。
Go to PDFnoted.com, click ‘OCR Scan Recognition’. Drag & drop or browse to upload your scanned PDF (supports A4 size; 300+ DPI recommended). PDFnoted auto-detects image-based PDFs and activates OCR instantly.
Select language(s) — e.g., ‘Simplified Chinese + English’ — then click ‘Start OCR’. PDFnoted’s AI model processes each page in 3–8 seconds, highlighting recognized text and confidence scores in real time.
After OCR completes, click ‘Export as Text’ for clean TXT (preserves paragraphs), or ‘Export as Searchable PDF’ to generate a new PDF with invisible text layer—keeps original layout while enabling search, copy & paste. All processing happens locally in your browser.
目前仅支持印刷体文字识别（如打印文档、清晰扫描件），暂不支持手写体；建议扫描分辨率≥300 DPI、对比度充足。模糊或倾斜严重的页面识别准确率会下降。
不能直接编辑PDF本身，但「导出为文本」可一键生成TXT，粘贴至Word后即可自由排版；若需保留格式，建议用「可搜索PDF」配合Word的‘插入→对象→PDF文件’功能提取内容。
不会。PDFnoted 在浏览器内完成全部OCR运算（WebAssembly加速），原始文件与识别结果均不离开您的设备，无数据上传，保障隐私与合规性。
PDFnoted OCR supports printed text only (e.g., typed documents, clear scans), not handwriting. For best results: scan at ≥300 DPI with high contrast. Accuracy drops significantly on skewed or low-resolution pages.
You can’t edit the PDF file directly in Word—but ‘Export as Text’ gives you clean TXT for pasting into Word. To retain layout, use Word’s ‘Insert → Object → PDF File’ to extract content from the searchable PDF.
No. PDFnoted runs OCR entirely in your browser using WebAssembly—no files are uploaded. Both input and output stay on your device, ensuring privacy and GDPR/CCPA compliance.