Scan to PDF

Place the page on a flat surface with a clear border around every edge. Use steady, even light and move lamps until glare no longer covers printed text, signatures, stamps, or photographs. A dark background helps the page boundary remain visible when the paper is white. Smooth folds where possible and remove clips that lift a corner. The camera captures the visible frame as a rectangular image. It does not detect the page boundary, correct perspective, remove fingers, or reconstruct detail hidden by a shadow.

Clean the camera lens before starting. Hold the device parallel to the page rather than aiming from one side, because an angled capture makes one edge appear smaller than the other. Fill most of the preview while leaving a narrow border for later checking. Use the rear camera when the device offers one. It usually has better focus and resolution than the front camera. Keep sensitive pages away from reflective windows, video calls, and other applications that may also use the camera.

Select Open camera and respond to the browser permission prompt. PWRKIT requests video only and does not request microphone audio. The browser shows a camera indicator while the stream is active. If access is denied, change the site permission in the browser or import existing JPG and PNG photos. A managed device may block camera access through operating system or workplace policy. The tool cannot override those controls, and it does not display a fake camera interface when no stream is available.

The camera stream stays attached to the current tab. Select Close camera when capture is finished, or reset the workbench to stop every active track. Closing the tab also ends the stream through normal browser cleanup. Switching cameras stops the current track before asking for the other facing mode. Some desktop computers and single-camera phones ignore that preference. Check the live preview after every switch instead of assuming the requested lens was selected.

Wait until the preview is sharp, then select Capture page. The current video frame is encoded as a JPG at 92 percent quality and added to the same ordered list used for imported files. The capture receives a page number in its filename for review, but the displayed list controls the PDF sequence. The live stream remains open so the next page can be photographed without another permission prompt. Capturing the same page twice creates two separate list entries.

Review every item before export. Look for soft focus, motion blur, glare, clipped corners, strong color casts, and text that becomes unreadable when zoomed. Remove a weak capture and take it again while the paper is still available. The tool does not automatically sharpen, deskew, crop, clean a background, or enhance handwriting. Those operations can alter evidence or make small characters less accurate, so the current workflow keeps the captured pixels and leaves visual judgment to the person preparing the document.

Camera capture and file import can be mixed in one job. Drop JPG or PNG files into the workbench, choose them from the device, or paste image files from the clipboard. The validator checks the extension and actual signature before adding a candidate. Renaming a PDF, WebP, or text file to .jpg does not make it a supported image. Up to 50 pages and 50 MB of combined stored input are accepted. The limit applies to captured frames and imported files together.

Existing photos are useful when pages were photographed earlier, when a desktop browser has no camera, or when a separate camera application offers controls needed for the task. Import the original photos rather than screenshots of a photo viewer. Screenshots can add controls, reduce resolution, or change color. If the device stores a photo as HEIC, export a JPG or PNG copy first. The scan workbench only lists formats that its PDF engine has verified.

One list item becomes one PDF page. Use the up and down controls to move pages until the sequence matches the paper document. Check front and back sides, inserts, appendices, and any deliberately blank page. A camera filename can help identify when an item was captured, but it does not override the list position. Removing an item excludes it from the next export. Adding another capture appends it to the end, so inspect the full sequence again after a retake.

For longer records, compare the list with printed page numbers or a written checklist. A missing middle page is easier to find before the paper stack is put away. The PDF builder does not infer order from timestamps, optical text, or filenames. It also does not merge two photographs into one page. If a wide sheet needs two captures, decide whether the recipient expects two consecutive pages or a separately prepared stitched image.

Match each image creates a PDF page around the source dimensions plus the chosen margin. A4 and US Letter create consistent paper boxes. Portrait suits most letters, forms, and receipts, while landscape can suit wide tables or boards. The selected paper setting applies to every item. Use separate jobs when different groups need different paper orientation. PDF points define the page box, with 72 points equal to one inch. The margin control accepts zero through 100 points on every side.

Contain keeps the complete photograph visible and centers it inside a fixed paper page. It is the safer choice when writing, totals, signatures, or page numbers sit near an edge. Cover fills the available area and center crops the overflow. That can suit full-page photographs, but it can also remove document content. The tool has no automatic edge detector or per-page crop handle. Prepare a better crop before import when the fixed center crop is not acceptable.

The camera preview, captured JPG frames, imported files, and generated PDF remain in the current browser tab. Selecting Create PDF loads the PDF builder after interaction, embeds each validated image, and saves a separate download. No PWRKIT conversion worker receives the page images. Analytics records the tool operation, processor label, broad file type, count, duration, and size bucket. It does not include camera frames, document text, filenames, passwords, or PDF bytes.

Local processing avoids a server upload, but the source and result still exist on the device. Downloads may enter synchronized folders, backups, browser history, antivirus inspection, endpoint monitoring, or shared storage. Follow the handling rules that apply to the paper record. Save to an approved location, inspect the result, close the camera, and close the tab when finished. The site cannot erase copies already created by the device, operating system, browser, or storage provider.

Camera frames are stored as JPG with a 0.92 browser quality setting. Imported JPG bytes remain JPG and imported PNG bytes remain PNG inside the new PDF. The builder does not apply another photo enhancement pass. Quality depends on the lens, focus, light, motion, source resolution, and browser encoding. A large page box cannot add missing detail. Zoom into small print after export and compare names, account numbers, dates, decimal points, and handwritten marks with the paper original.

The PDF size is usually close to the combined images plus document overhead. Many high-resolution captures can create a large result even when the paper pages look simple. Reduce image dimensions before import or use Compress PDF after export if the delivery system has a strict limit. Compare the compressed copy with the first PDF at normal size and high zoom. Keep the clearer version until the recipient has accepted the document.

Printed words in a camera frame remain pixels. This tool does not create selectable text, run optical character recognition, assign reading order, tag headings, describe images, set a document language, or add form fields. Search, copy, and screen-reader access therefore differ from a source document with a text layer. If a submission requires OCR, run a reviewed OCR workflow after capture and compare every extracted name, number, table cell, and punctuation mark with the page.

OCR alone does not make a PDF accessible. Multi-column pages, tables, diagrams, handwritten notes, signatures, and stamps require structure and descriptions that an image capture cannot supply. Official publication may also require a document title, language, logical reading order, bookmarks, and tagged form controls. Use an accessibility editor and the recipient's validation process for those requirements. The scan tool packages page images and does not claim to produce a tagged archival record.

Open the downloaded file in a separate PDF reader. Confirm the filename, page count, sequence, orientation, paper size, margins, focus, color, and edge visibility. Inspect the first and last page, then sample pages from the middle. Zoom into the smallest important text. Compare totals and identifying numbers with the original paper. If a page was retaken, make sure the rejected version is not still present. A successful download proves that a PDF was created, not that the document is complete or acceptable.

Test the actual delivery path when possible. Email services, records portals, banks, schools, printers, and signature systems may enforce their own file-size, PDF-version, page-size, encryption, or image-quality rules. Keep the paper and captured source files until acceptance is confirmed. If the recipient rejects the result, note the exact message before changing settings. Rebuilding blindly can reduce quality without addressing the stated requirement.