Images and PDF

JPG to PDF: arrange images, choose page layouts and check the result

·

To convert images to PDF, choose your JPG, PNG or still WebP files, arrange their order and set the page layout before creating the document. ToolMellow processes the images on your device and produces one images.pdf, or a separate PDF for each image. Each image becomes one page.

Start with the recipient’s requirements: one document or individual files, a paper size or pages shaped like the images, and an acceptable file size. Then check the actual download. Converting a photograph to PDF packages its pixels into a document; it does not add searchable text or recover missing detail.

JPG to PDF

Choose a combined document or individual image pages

A combined PDF suits a sequence that must stay together: a receipt bundle, a short illustrated report or several scanned pages. Separate PDFs suit items that need individual filenames or separate submission fields. In both modes, one selected image becomes one PDF page; the tool does not place several images on a single page.

The name JPG to PDF describes the main task, but valid PNG and still WebP images also work. You can mix those formats in the same queue. HEIC, GIF and existing PDF documents are outside this route; convert a HEIC photo to JPG first, or use Merge PDF for existing documents.

Choose a combined document or individual image pages
What you needChooseWhat you receive
One ordered documentOne combined PDFA single images.pdf with one page per image
One document per imageA separate PDF for each imageNumbered single-page PDFs
A batch of separate resultsSeparate output, then Download all as ZIPtoolmellow-pdfs.zip containing the PDFs
Combine existing PDF pagesMerge PDFA separate document workflow, rather than image conversion

Follow a practical image-to-PDF workflow

Prepare readable source images before starting. Check that every scan includes the complete page, that photographs face the correct way and that confidential details are suitable for the intended recipient. The conversion settings affect presentation and encoding; they cannot repair an unreadable original.

  • Choose your images or drop them into the workspace. Additional selections append to the current queue.
  • Move, sort or reverse the files into the intended page order, and rotate individual images when necessary.
  • Choose combined or separate output, a page size, orientation, margins and image placement.
  • Choose JPG or PNG image encoding. For JPG, set quality and the background used for transparent pixels.
  • Wait for image previews to settle, examine the layout illustration and select Create PDF.
  • Download the actual result, inspect every page and use Download all as ZIP when you need multiple separate PDFs.

iLovePDF: a guide to converting JPG images to PDF

Image order and rotation determine the page sequence

The visible queue is the output order. Move an image up or down for an exact position, use Sort by filename for a filename-based sequence, or use Reverse order to invert the queue. Filename sorting compares names with numeric ordering, so numbered names are useful, but inspect the final list rather than assuming its order from your file picker.

Each image has its own Rotate 90° control. Rotation cycles through quarter turns and stays associated with that image when you move, sort, reverse or remove other files. The browser first honors the source image’s orientation metadata; the selected clockwise turn is added afterward. Changing the queue or settings clears the previous result, so create a new PDF after an edit.

MDN: createImageBitmap orientation and decoding options

Choose A4, US Letter or image-sized pages

Use the paper size requested by the person or system receiving the document. A4 and US Letter provide a consistent paper format; each image is positioned within that page’s margins. Fit image creates a different page footprint from each rotated image’s pixel dimensions, using a fixed 96 pixels per inch.

For an image-sized page, the calculation is page points = image pixels × 0.75. A 1200 × 800 pixel image therefore produces a 900 × 600 point page before margins affect placement. The page size comes from this calculation, rather than a DPI tag in the source file. Choose zero margins when you want a page that follows the image’s proportions without an inset.

Choose A4, US Letter or image-sized pages
Page choiceBase sizeUseful when
A4595.28 × 841.89 PDF points before orientationThe recipient requests A4 paper
US Letter612 × 792 PDF points before orientationThe recipient requests US Letter paper
Fit image (96 PPI)Rotated pixel width and height multiplied by 0.75The page should follow each image’s aspect ratio

pdf-lib PDFDocument API: page sizes and image embedding

Choose page orientation after checking the rotated image

Portrait and Landscape set the paper orientation for A4 or Letter. Automatic matches each image after its selected rotation: a wider-than-tall image gets a landscape paper page, while a taller or square image gets portrait. A combined document can therefore contain both orientations when Automatic is selected.

Fit image disables the Orientation control because the page already derives its shape from the rotated image. To change that page’s shape, rotate the image itself. Changing a paper page to landscape changes the available physical space; it does not add pixels or make a low-resolution photograph sharper.

Margins must leave room inside the chosen page

Margin is measured in millimetres and accepts values from 0 to 50. The margin is applied on every side. It must leave a positive printable width and height, so the same margin can work on A4 yet fail on a small image-sized page. The default is 10 mm, which can consume all the available space on a very small image.

If the tool reports that margins leave no room, reduce the margin or choose A4 or Letter. Margins on Fit image shrink the area available for the image; they do not enlarge the page outside its calculated footprint. A zero margin also does not guarantee edge-to-edge printing, because your printer or print-dialog settings can add their own restrictions.

Fit inside shows the whole image; Fill crops visible edges

Fit inside keeps the image’s proportions and centers the whole image within the margins. When its shape differs from the printable area, unused space remains. Fill printable area scales the image to cover that area and clips the centered overflow. Choose Fit inside for complete receipts, certificates and scans whose edges matter.

Fill can suit a photograph when losing part of the edges is acceptable. Examine the layout before creating the PDF and check the download afterward. The clipping affects what the PDF displays; hidden image pixels can remain embedded in the document. Use an appropriate image-editing or redaction workflow when content must actually be removed.

pdf-lib PDFDocument API: page sizes and image embedding

Choose JPG encoding for adjustable quality or PNG for alpha

The default JPG encoding uses quality 90 and a white background. The quality slider ranges from 10 to 100 and controls the browser’s lossy encoder. A lower setting can reduce the embedded image data but can also introduce visible artefacts, especially around lettering. Quality 90 is an encoder setting, rather than a measurement that 90 percent of the original visual quality remains.

PNG retains the rendered image pixels and alpha without a lossy quality setting, so the quality slider is disabled. It can produce a larger PDF, especially for photographs. With JPG encoding, the chosen background fills transparent image pixels; it does not replace an opaque background or set the color of the page margins. PNG alpha also does not guarantee that every PDF viewer shows a transparent page background.

Choose JPG encoding for adjustable quality or PNG for alpha
DecisionUseful choiceCheck afterward
Photo with a file-size requirementJPG, then adjust qualityFine detail and the actual PDF size
Diagram or image needing alphaPNG encodingTransparency appearance in the intended viewer
Transparent image exported as JPGChoose a suitable JPG backgroundColors at transparent edges
Small lettering or line detailCompare suitable JPG and PNG resultsReadability at the intended viewing or print size

MDN: createImageBitmap orientation and decoding optionsMDN: OffscreenCanvas image encoding with convertToBlob

Use the layout preview for placement, then inspect the PDF

The Preview image selector shows a chosen source image within the planned page. The dashed guide marks the printable area and does not appear in the exported PDF. Use it to spot an unexpected rotation, excess blank space or the edge cropping caused by Fill.

This is a small layout illustration built from a reduced source preview. It does not render the finished PDF bytes or establish full-resolution image quality. After downloading, open the document in the viewer you intend to use and examine every page, including its size, order, crop and fine detail.

Image pixels and paper dimensions are different quantities. Placing the same image over a larger physical area reduces its effective pixels per inch; placing it over a smaller area increases that density. The relationship is effective PPI = image pixels ÷ placed size in inches. Page layout controls physical placement while retaining the rotated image’s pixel dimensions.

Turning a JPG into a PDF cannot restore detail that the camera, scanner or earlier compression has lost. The route decodes and re-encodes the image rather than embedding the original JPG bytes unchanged. Keep the original, use a better source when needed and inspect a print sample if the result is intended for paper.

MDN: OffscreenCanvas image encoding with convertToBlob

Image pages do not add searchable OCR text

A photograph of a document remains an image inside the PDF. This operation does not add an OCR text layer, editable paragraphs, document headings or accessibility tags. Choosing PNG or a higher JPG quality setting affects image encoding, not text recognition.

If you need text from a scan, use a separate OCR workflow and review its recognition errors. ToolMellow’s Image to text tool is a separate option with its own language and input limits. OCR and scan-enhancement features described by other PDF products belong to those products and should not be assumed to exist in this conversion.

Adobe Acrobat: scanning documents and recognizing text

Local conversion still needs the website’s code and assets

Selected image contents are processed on your device by the local image worker. You do not need an account to create the PDF. Loading the page, worker and libraries can still make network requests; local file processing is not a promise that the website is an installed offline application.

The conversion does not promise an unchanged original file, removal of every metadata field, secure redaction or a managed print-color profile. Original image metadata is not explicitly copied through this image-to-PDF route, but new encoded-image and PDF metadata can exist. Keep your originals and assess the actual output for the purpose in which it will be shared.

MDN: createImageBitmap orientation and decoding optionsMDN: OffscreenCanvas image encoding with convertToBlob

Distinguish source limits from encoded images and final PDFs

A file can meet its individual limit while the whole job exceeds a different limit. The tool checks both the selected source files and the generated data. MB here uses decimal bytes. Keep the source count, source dimensions and combined source size within the bounds before expecting conversion to succeed.

The cumulative encoded-image limit measures image buffers before PDF embedding. The 50 MB limit applies to each final serialized PDF. In separate mode, the PDFs are not required to total 50 MB, and the 100 MB encoded-image limit is not a 100 MB limit on the sum of final PDFs. Output size also depends on encoding, image content and document overhead; quality settings do not guarantee a target size.

Distinguish source limits from encoded images and final PDFs
LimitMaximumScope
Selected images20One conversion job
Source file size25000000 bytes (25 MB)Each image
Combined source size100000000 bytes (100 MB)All selected files together
Source dimensions24000000 pixels and 12000 pixels per sideEach image must satisfy both bounds
Encoded image data100000000 bytes (100 MB)Cumulative image buffers across the job
Final PDF size50000000 bytes (50 MB)The combined PDF or each separate PDF
Conversion deadline60 secondsThe whole worker job, including startup

Recover from invalid images, large results or cancellation

An extension is not enough to make a file valid. If a source cannot be read, export a fresh JPG, PNG or still WebP from an image application and try again. Animated PNG and animated WebP are rejected; export a still frame first. Oversized dimensions are rejected before ordinary decoding when the header exposes them, and reducing PDF page size does not bypass those source limits.

If a later image fails, the job returns an error without a partially completed result list. Remove or replace the failing source, adjust unsuitable settings and create the PDFs again. For output-size or time limits, use fewer or smaller images or try suitable JPG encoding. Cancel stops the worker and retains the original selected inputs; no partial completed download is returned. Clear files removes the queue and results but retains the current PDF settings.

Download the result and continue only when useful

Combined output is named images.pdf. Separate results use a numbered prefix and a safe version of the source name, such as 01-page.pdf; the numbering keeps duplicate source basenames distinct. Use each Download button for an individual PDF, or Download all as ZIP for multiple results. Check the recipient’s size requirement against the actual download rather than the source image sizes.

Each result can continue to Merge PDF, Compress PDF or Review text for Word as a separate workflow. The transfer is explicit, temporary and held in memory in the same tab; it is not saved in a URL or persistent browser storage, and reloading loses it. Passing image pages to a text-review tool does not create OCR text. Download a copy you want to keep before leaving the session.

Common questions

Can I combine JPG, PNG and WebP images in one PDF?

Yes. Valid JPG/JPEG, PNG and still WebP images can share the same queue. Choose One combined PDF for one page per image in the visible order. Animated PNG and WebP are unsupported, and HEIC or existing PDFs need a different workflow.

How do I save one PDF per image?

Choose A separate PDF for each image in PDF output, then create the results. Each PDF has one page and a numbered safe filename. Download the files individually or use Download all as ZIP when multiple results are available.

Why are the edges of my image missing from the PDF?

Fill printable area scales the image to cover the space inside the margins and clips centered overflow. Choose Fit inside to show the whole image. The clipping is a display operation and should not be treated as secure removal of hidden image content.

Why does an image-sized page reject my margins?

The page footprint is calculated at 96 PPI from the rotated image. Margins must leave positive space inside that footprint. A small page can be too small for the default 10 mm margins; reduce the margin, use zero when appropriate or choose A4 or Letter.

Will converting a JPG to PDF improve its image quality?

No. It cannot recover missing source detail. The image is decoded and re-encoded, and its physical placement affects effective pixel density. Compare the actual result at the intended viewing or print size, keeping the original image.

What are the image and PDF size limits?

A job supports 20 images, 25000000 bytes each and 100000000 source bytes combined. Each source is limited to 24000000 pixels and 12000 pixels per side. Encoded image buffers have a separate 100000000-byte cumulative cap; every resulting PDF has a 50000000-byte cap.

Does JPG to PDF make scanned text searchable?

No. It creates image pages without adding an OCR text layer. Text recognition is a separate task; review any OCR result rather than assuming it is an accurate or accessible transcription of the scan.

Are my images uploaded during conversion?

The selected image contents are processed on your device by the local worker. Website code and library assets can still load over the network. Local processing also does not certify that the resulting document has no metadata.

How can I make the resulting PDF smaller?

Try suitable JPG encoding, compare lower quality settings or resize overly large source images in a separate workflow. Use fewer images when appropriate. Check the actual PDF size and readability; reducing paper dimensions alone is not a pixel-downsampling control.

Does PNG encoding preserve transparency?

It retains rendered alpha in the embedded image, and the quality slider is disabled. The PDF page may still appear white in a viewer. With JPG encoding, transparent image pixels use the selected JPG background instead.

What does Automatic orientation use?

For A4 and Letter it checks each image after the selected rotation. A wider image receives landscape; a taller or square image receives portrait. Fit image disables the paper-orientation control because its page dimensions already follow the image.

Does the layout preview show final PDF image quality?

No. It is a reduced placement illustration with a dashed printable-area guide. Download and open the PDF to inspect fine detail, clipping, dimensions and appearance in the intended viewer. The dashed guide is not exported.

Does cancelling change my original images?

Cancel terminates the conversion worker and keeps the selected original inputs. A failed or cancelled job does not return a partial completed result list. You can adjust the job and try again, or clear the queue.

Why did the images disappear after a workflow handoff and reload?

The handoff is held once in memory in the same tab, without persistent storage or URL parameters. Reloading clears it. Save the PDF you want to keep and choose the file again if you need to resume later.

Primary sources and technical references

Put it into practice.

Related guides