Scan a Document to PDF With Your Camera
Scanning a document to PDF with a camera means photographing each paper page and packing the captures into one multi-page PDF, using the camera your browser already has access to instead of a flatbed scanner. This tool shows a live viewfinder, lets you capture page after page — or pick existing photos through the file picker instead — and assembles everything into an A4 PDF on your device. Camera frames are processed locally and never leave your device.
How it works
The viewfinder is powered by getUserMedia, the standard browser API for camera access. When you allow access, the browser hands the page a live MediaStream that is attached to a video element — the stream flows directly from the camera hardware to your screen and is never routed through any server. On phones the tool requests the rear (environment-facing) camera at the highest resolution the device offers, because document legibility depends almost entirely on pixels per letter.
Pressing capture copies the current video frame onto an offscreen canvas — a one-frame snapshot, not a recording — and encodes it as JPEG. JPEG is the right codec here for two reasons: photographs of paper compress far better with it than with lossless formats, and PDF supports JPEG natively through its DCTDecode filter, so each capture can later be embedded without being decoded and re-encoded a second time. Every capture is added to a thumbnail list where you can review it, retake a blurry page, or delete a duplicate before committing to the document. If you would rather not grant camera access at all, the same list accepts photos chosen through the ordinary file picker.
The assembly step creates one A4 page (595 × 842 points) per capture with pdf-lib and draws the image scaled to fit inside the page while preserving its aspect ratio — a landscape-shaped capture is not stretched to portrait, it is fitted with white margins. Be aware of what this tool intentionally does not do: there is no perspective de-skewing, no shadow removal, and no OCR, so the output is a faithful photographic record rather than a cleaned "scan," and its text is not selectable. For sharpness, hold the camera parallel to the page and fill the frame; for searchable text, run the finished file through an OCR tool afterwards.
Worked example: an 8-page contract scanned on a phone
Using a mid-range phone with a 12-megapixel rear camera, capturing all 8 pages of a printed contract took just over two minutes, including one retake for a page shot at an angle. Each capture encoded to a JPEG between 780 KB and 1.1 MB, and the assembled A4 PDF came to 7.6 MB — small enough for most email limits but large enough that the test continued with a pass through Compress PDF, which brought it down to 2.1 MB at a quality still comfortably readable on screen.
A second run on a laptop with a 720p webcam illustrates the resolution floor: the same first page captured at 1280 × 720 produced only about 55 pixels per line of 11-point text, and small print was legible but soft. Phone cameras capture 3 to 4 times more linear resolution, which is why the phone is the better scanner whenever both are available.
Frequently asked questions
Are my camera pictures uploaded to a server?
No. The video stream goes straight from the camera to your browser tab via the getUserMedia API, captures are held in local memory, and the PDF is assembled on your device. Nothing is transmitted over the network at any point.
Can I use this without granting camera access?
Yes. The file picker option accepts photos you have already taken, so you can photograph pages with your normal camera app and assemble them here. The result is the same multi-page A4 PDF.
Why does the browser ask permission every time or block the camera?
Browsers only expose cameras on secure (HTTPS) pages and always require an explicit permission grant, which you may have previously declined. Check the camera icon in the address bar to reset the permission, and note that another app already using the camera can also block access.
Will the scanned PDF text be searchable?
No — a camera scan is a photograph, so the PDF contains images of text, not text itself. Making it searchable requires OCR, which this tool does not perform; run the file through an OCR program if you need selectable text.
How large is a PDF made from camera captures?
Expect roughly 0.5 to 1.5 MB per page from a modern phone camera after JPEG encoding, so a 10-page scan usually lands between 5 and 15 MB. If that is too big to email, the Compress PDF tool can re-encode the pages at a lower quality.
Does the tool straighten pages or remove shadows like a scanner app?
No. Captures are placed on the page exactly as photographed, with no perspective correction, cropping, or contrast enhancement. You get the best results by shooting straight down onto the page in even light.