Why people convert images to PDF in the first place
JPG, PNG and HEIC are great containers for a single image. They’re terrible for a multi-page document. The moment you need to send a stack of receipts, a passport copy with three pages, a scanned contract, or a portfolio of product shots, PDF wins on three axes: page order is preserved, every recipient can open it without an image viewer, and the whole document is one drop-in attachment instead of a zipfile or a noisy email thread.
Done well, JPG-to-PDF preserves every pixel of every image, fixes phone-camera rotation automatically, lets you reorder pages without re-photographing anything, and ships a file that prints crisply at A4 or Letter. Done poorly — by an engine that silently downscales, ignores EXIF, or uploads your ID scan to a third-party server — you get a small, blurry, sideways, privacy-leaked PDF that costs you a re-submission.
How page sizing actually works
Every PDF page has a size measured in PostScript points (1 pt = 1/72 inch). A4 is 595 × 842 pt, US Letter is 612 × 792 pt, A3 is 842 × 1191 pt. When you pick "Auto", the tool sets the page size to match the image at 72 DPI — your 4000 × 3000 photo becomes a page that prints at 55 × 41 inches, which is enormous. That’s fine for a single-image quick share, terrible for a multi-image document.
- A4 / Letter — the right pick for documents you’ll print or share on a desk. Images are scaled to fit with the aspect ratio preserved.
- A3 / Tabloid — large-format design proofs, posters, blueprints.
- A5 — booklets, half-page handouts, mobile-first reading.
- Original — keep the page the exact pixel dimensions of the image (good for portfolios, bad for printing).
- Auto — best for one-off shares; the page is sized to the image.
Fit modes explained
Page size and image size rarely match exactly. The "Image fit" option decides what to do:
- Fit to page — scale uniformly so the whole image is visible (the default). Adds whitespace on one axis if the aspect ratios differ.
- Fill page — scale uniformly so the page is fully covered. Crops the image on one axis. Great for full-bleed photo books.
- Stretch — scale non-uniformly to match the page exactly. Distorts the image; use only for diagrams that don’t care about aspect ratio.
- Center at native size — place the image at its native pixel size, centred on the page. Adds whitespace if the image is smaller; clips if larger.
The EXIF rotation trap
This is the single biggest "why is my PDF sideways?" issue. Phone cameras shoot at the sensor’s native landscape orientation and store a flag — EXIF Orientation = 6 — that says "rotate 90° CW when displaying". Photo viewers honour the flag. PDF embedders almost never do, unless they’re explicitly written to. Our tool reads the EXIF tag, pre-rotates the pixel buffer, then embeds the upright image. Portrait phone photos come out portrait in the PDF.
How transparency is handled
PNG, WEBP and HEIC can carry an alpha channel. JPEG cannot. If you re-encode a transparent PNG as JPEG without compositing, you get black blocks where transparency used to be. We solve it the right way: every transparent image is rendered onto a solid background (white by default; black or custom on request) before the JPEG encoder ever sees it. Or you can opt out — pick "Preserve transparency" and we embed the PNG stream directly into the PDF, alpha and all, with no re-encoding.
Compression: the size-vs-fidelity dial
The "Compression / quality" slider controls the JPEG quality factor used when re-encoding. 0.92 is the sweet spot for photos — visually indistinguishable from source. Drop to 0.85 for web sharing, 0.75 for email-size, or push to 1.00 for archival output. If you disable "Re-encode to JPEG", we embed the original image bytes directly: smallest possible quality loss, but typically a much larger PDF, especially for PNG sources.
Memory discipline: why most converters crash on big batches
A naive converter decodes every image into a Canvas, holds them all in memory, then assembles the PDF in one pass. For a 200-photo album at 6000 × 4000 RGBA, that’s ~12 GB of RAM — the browser tab dies. Our pipeline streams: decode image → optionally re-encode → embed into pdf-lib → drop the canvas and pixel buffer → garbage-collect → move to next. Steady-state memory stays under 1.5 GB for batches up to ~1,000 images.
Privacy: the part no one talks about
"Free" online JPG-to-PDF converters are, almost without exception, server-side. Your ID scan, signed contract, medical photo or financial document is uploaded to someone else’s server, parsed there, cached on disk, and — in many cases — indexed for ad targeting or training data. For GDPR-, DPDP- and HIPAA-bound documents, that single upload is the breach.
Browser-based conversion sidesteps the issue entirely. Open your network tab while converting on this page — you’ll see exactly zero outbound requests carrying your image data. Decoding, EXIF handling, OCR (optional), JPEG re-encoding and PDF assembly all happen in JavaScript and Canvas, on your CPU.
The full image-to-PDF workflow professionals actually use
- Drop or pick all the images in one go (mix JPG, PNG, HEIC, TIFF freely).
- Reorder thumbnails by drag-and-drop. Rotate any sideways pages. Delete the duds.
- Pick a page size — A4 for documents, Original for portfolios, Auto for quick shares.
- Pick fit mode — "Fit to page" is the safe default; "Fill" for photo books; "Center at native size" for technical exhibits.
- If photos are dark or scanned, turn on "Auto-enhance".
- If the PDF needs to be searchable, turn on "Add OCR text layer".
- Convert. Click Download. Done.
Case study: a property advocate cutting filing time in half
Problem: A property law practice in Pune was receiving photographed sale-deed pages and ID proofs from clients over WhatsApp — typically 8 to 30 phone photos per file, in mixed orientation, mixed resolution, with HEIC files from iPhone clients. The paralegal team spent ~25 minutes per file rotating, ordering and exporting to PDF using a desktop tool.
Process: The team switched to this in-browser converter. Drag photos from the WhatsApp download folder, drag thumbnails into deed order, click Convert, attach the PDF to the case file. Average time per file: under three minutes. Because images never leave the laptop, the firm’s client-confidentiality clause is satisfied without a vendor NDA.
Result: Per-file turnaround dropped from 25 minutes to under 3 — an 8× speed-up. Across a 60-case month, that returns ~22 paralegal hours to high-value work.
Tips for the cleanest possible PDF
- Match page size to use-case. A4 for documents, Original for portfolios, Auto for one-off image shares.
- Quality 0.85–0.92 is the sweet spot for photos. Drop to 0.75 for web; push to 1.00 only for archival.
- Pick the right background if any image has transparency — white is safest, black is great for slide decks.
- Leave EXIF auto-rotation on unless you’ve already pre-rotated the originals.
- For scans, turn on Auto-enhance; for screenshots, leave it off (it can flatten clean UI colours).
How we measure conversion quality
Our internal QA suite runs 1,000 reference batches through the converter on every release — phone albums, scanned receipts, ID photos, product shots, blueprints, screenshots and mixed-format dumps. Each output is scored on visual fidelity (vs. source pixel hash), orientation correctness (EXIF compliance), color accuracy (ΔE2000) and assembly success (no missing pages, no embedded-stream errors). The current build sits at 99.6% visual fidelity, 100% orientation correctness and 99.95% assembly success across the suite.