# OCR PDF — Technical Tool Documentation > **Summary:** Recognize text in scanned PDF pages with browser-based Tesseract OCR. > **Tool Route:** `/tools/ocr.html` > **Visual Guide Route:** `/guides/ocr-pdf/` > **Category:** Extract Content > **Processing Architecture:** Hybrid: Text extraction is 100% local in WebAssembly Worker. Network access is required once to download language traineddata from CDN. --- ## 1. What the Tool Does OCR PDF renders selected pages and sends the images through the locally loaded Tesseract.js worker. It returns recognized text rather than a searchable PDF. **Requirements & Constraints:** A network connection may be needed to load PDF.js, Tesseract, and language data. Select a supported OCR language and page range. --- ## 2. When to Use It Use **OCR PDF** whenever you need to perform client-side extract content operations without uploading confidential records to third-party cloud servers. **Ideal Scenarios:** - Working on personal identity records, banking statements, medical filings, or NDAs requiring 100% local confidentiality. - Fast offline or low-bandwidth processing directly in your web browser. - Universal cross-platform document operations on Windows, macOS, Linux, Android, or iOS. --- ## 3. Key Implemented Features - **Engine Architecture:** `pdf.js (canvas rendering) + Tesseract.js (Optical Character Recognition)` - **Zero Server Uploads:** Processing executes directly within client browser memory. - **Instant Processing:** Fast execution powered by WebAssembly and native HTML5 APIs. - **Review Previews:** Visual feedback and progress indicators confirm results prior to saving. - **No Sign-up / No Watermarks:** Full functionality with zero paywalls or forced branding. --- ## 4. Supported Input Files - **Primary Input:** Scanned or image-only PDF document (.pdf) - **Validation:** Files are validated in the browser prior to parsing. --- ## 5. Supported Output Files - **Output Format:** Searchable extracted text (.txt) and on-screen copyable text - **MIME Types:** `application/pdf` (or appropriate image/text/zip formats). --- ## 6. File Size & Memory Limits - **Recommended Bounds:** Client RAM bound (~20 pages per batch recommended) - **Technical Context:** Modern browser tabs enforce memory quotas (typically 1.5 GB to 2 GB per tab). PDFHub handles documents comfortably within standard browser memory limits. For files exceeding 150 MB or 1,000 pages, splitting into smaller sections is recommended. --- ## 7. Step-by-Step Instructions 1. **Upload a scanned PDF.** 2. **Choose the recognition language and optional pages or ranges.** 3. **Select Extract Text and wait while each page is recognized.** 4. **Review the text, then copy it or download the TXT file.** --- ## 8. Required User Inputs & Controls - **File Upload Area (`#drop` / `#fin`):** Click to open native file browser or drag and drop supported files. - **Tool Options (`#opts`):** Dynamic settings panel configuring parameters, page ranges, or compression levels. - **Execution Button (`Extract Text with OCR (#goBtn)`):** Triggers client-side processing pipeline. - **Reset Button:** Clears loaded files and resets workspace to pristine state. --- ## 9. Processing Behavior & Execution Pipeline Renders PDF pages to high-resolution canvas images, feeds pixel matrices to Tesseract.js worker, performs character recognition against selected language traineddata, and compiles output text. --- ## 10. Expected Output & Download Verification - Upon successful execution, the interface triggers an automated browser download via `downloadBlob()` or opens a result review modal. - Users can verify file size, page count, and visual rendering before finalizing the download. --- ## 11. Practical Example Use Case Extracting text and tables from flat image-only PDF scans of historical books or paper invoices. --- ## 12. Common Errors & Troubleshooting - **Issue:** If language data cannot load, check connectivity and retry. - **Issue:** If the browser becomes slow, process fewer pages at a time. - **Avoid:** Expecting perfect recognition from skewed or low-resolution scans. - **Avoid:** Choosing the wrong language for the document. - **Tip:** Low resolution scans producing poor OCR accuracy. Solution: Ensure scans are at least 150–300 DPI. - **Tip:** Selecting the wrong language dictionary. Solution: Choose the matching language from the dropdown menu. --- ## 13. Privacy & Data Handling Architecture - **Local Processing:** Hybrid: Text extraction is 100% local in WebAssembly Worker. Network access is required once to download language traineddata from CDN. - **Browser Memory:** Document bytes reside temporarily in `ArrayBuffer` memory and are garbage collected when the browser tab is refreshed or closed. - **Cookies & Telemetry:** PDFHub sets zero tracking cookies, zero analytics beacons, and stores zero personal document data. --- ## 14. Frequently Asked Questions (FAQ) #### Q: Does OCR create a searchable PDF? **A:** No. It outputs recognized text and a text download. #### Q: Is OCR fully offline? **A:** Required libraries and language data may be loaded from CDNs. --- ## 15. Related PDFHub Tools - [PDF to Markdown Documentation](pdf-to-markdown.md) - [PDF to Word Documentation](pdf-to-word.md) - [AI Summarizer Documentation](summarize-pdf.md) --- ## 16. Links & Resources - **Launch Live Tool:** [Open OCR PDF in PDFHub](../../tools/ocr.html) - **Visual Tutorial:** [View OCR PDF Visual Guide](../../guides/ocr-pdf/) - **All Tools Master Reference:** [All Tools Reference](../all-tools-reference.md) - **Master Documentation Index:** [Documentation Index](../documentation-index.md) --- ## 17. Visual Walkthrough & Interface Screenshots ### Step 1: Open OCR PDF - **Interface Control:** `Upload Drop Zone (#drop)` - **What You See:** The drop area accepts scanned PDF documents or image files. - **What to Do:** Open OCR PDF to recognize and extract text from scanned documents or photos. - **Practical Pro-Tip:** OCR uses an embedded client-side neural recognition engine. ![OCR PDF Step 1 - Open OCR PDF](../images/ocr-pdf-step-1.png) ### Step 2: Upload Scanned Document - **Interface Control:** `File Browser (#fin)` - **What You See:** The document loads and total page count is detected. - **What to Do:** Select your scanned PDF containing images of printed text. - **Practical Pro-Tip:** Higher-resolution scans yield higher character recognition accuracy. ![OCR PDF Step 2 - Upload Scanned Document](../images/ocr-pdf-step-2.png) ### Step 3: Select Document Language - **Interface Control:** `Language Dropdown (#lang)` - **What You See:** The dropdown updates the language training model used for character matching. - **What to Do:** Choose the primary language of the text in your document (e.g. English, Spanish, German). - **Practical Pro-Tip:** Selecting the exact language dramatically improves accuracy for accented characters and punctuation. ![OCR PDF Step 3 - Select Document Language](../images/ocr-pdf-step-3.png) ### Step 4: Set Page Range - **Interface Control:** `Page Range (#pgrange)` - **What You See:** The input confirms your selected page limits. - **What to Do:** Specify which pages to process, or leave blank to process all pages. - **Practical Pro-Tip:** For large documents, testing OCR on pages 1-2 first allows you to verify recognition quality quickly. ![OCR PDF Step 4 - Set Page Range](../images/ocr-pdf-step-4.png) ### Step 5: Click Recognize Text - **Interface Control:** `OCR Action Button (#goBtn)` - **What You See:** The optical character recognition engine analyzes letter glyphs and words page by page. - **What to Do:** Click the primary "Recognize Text" button. - **Practical Pro-Tip:** A progress readout shows real-time page recognition status. ![OCR PDF Step 5 - Click Recognize Text](../images/ocr-pdf-step-4.png) ### Step 6: Review Recognized Text - **Interface Control:** `Text Result Area (#ocrResult)` - **What You See:** The pane displays selectable, editable text organized by paragraphs and lines. - **What to Do:** Inspect the extracted text rendered directly inside the editable result pane. - **Practical Pro-Tip:** You can make inline edits directly in the text box before exporting. ![OCR PDF Step 6 - Review Recognized Text](../images/ocr-pdf-step-4.png) ### Step 7: Copy or Download Output - **Interface Control:** `Action Buttons (copyText, downloadTxt)` - **What You See:** The copy confirmation confirms text is on your clipboard ready to paste into any editor. - **What to Do:** Click "Copy Text" to copy directly to clipboard or click "Download .txt" to save a file. - **Practical Pro-Tip:** No scan data ever leaves your device during the entire OCR process. ![OCR PDF Step 7 - Copy or Download Output](../images/ocr-pdf-step-4.png)