Black & White Scan Converter

This tool reduces a document image to two tones — pure black and pure white, with nothing in between. That is exactly what you want when the content is text, line art, a form, or a receipt and you care about legibility over fidelity. Stripping the grays removes coffee-colored paper, faint shadows, and the muddy mid-tones that make a phone photo of a page look like a phone photo of a page.

It fits a handful of concrete jobs. For archiving, a 1-bit-style page compresses far smaller than a full-color JPG, which keeps a folder of filed paperwork or a ZIP of scanned records lean. For OCR prep, feeding crisp black text on clean white to a text-recognition step often behaves more predictably than handing it a shadowed, low-contrast original. For office and admin work, it makes a tidy photocopy-style version of a contract page, an invoice, or a signed form before printing or attaching it to an email.

It is the wrong choice when tone matters. A portrait, a product shot, or any image with continuous shading will collapse into harsh blotches, because there is no gray left to carry the gradient. Reach for it when the subject is ink on paper, not when it is a photograph.

Everything runs in your browser. You drop a file onto the dropzone or pick one with the file selector — JPG, PNG, or WebP, up to 30 MB. The image is decoded onto a canvas locally; on this site it is not sent to a server, and once you close the tab nothing about it remains.

You then choose a threshold mode. Adaptive is the default and the better starting point for phone photos and uneven scans: it computes a separate cutoff for each small region of the page, so a shadow falling across one side does not turn that whole side black. Fixed mode gives you one global cutoff via a slider from 50 to 220 (default 128) — lower values push more pixels to black, higher values push more to white — and it shines when the original lighting is already flat and even. Press Convert to B&W, compare the original against the result in the before/after slider, and download a PNG named after your file with a -bw suffix.

Be clear on what it does not do. There is no clipboard paste, no batch queue, and no multi-page handling — it processes one image at a time. Output is always a single PNG; it does not produce a PDF, and despite the visual result being two-tone, the file is a standard PNG rather than a true 1-bit bitmap. For multi-page or PDF assembly you would chain it with a separate scan-to-PDF step.

Noise is the most common disappointment. Because every pixel is forced to either black or white, dust, JPEG compression artifacts, and sensor grain in a dim photo get amplified into scattered black speckles across the white background. If the original is noisy, clean or de-speckle it first; the threshold step has no middle gray to hide imperfections in.

Mode and slider mismatches cause the rest. Using fixed threshold on a photo with a shadow gradient will blow out the bright corner to solid white and crush the dark corner to solid black — that is the case adaptive mode exists for. In the other direction, a fixed value set too low can fill in thin strokes and bleed light text together, while too high can erase faint pencil or pale ink entirely. Adaptive mode is robust but not magic: very low-contrast originals (gray text on gray paper) can still drop characters, and large flat regions can pick up faint background texture.

Two structural limits to keep in mind: anything over 30 MB is rejected, and the conversion is destructive in the sense that lost gray detail cannot be recovered from the PNG afterward. Keep your original file so you can re-run with different settings rather than trying to rescue the output.

Use the before/after slider to read the page critically, not just glance at it. Confirm every line of text is intact and continuous — watch for letters that have broken apart, thin strokes that vanished, or words that merged into solid blobs. Scan the background for speckle, check that signatures and stamps survived, and verify that any faint annotations or carbon-copy text are still present. If something dropped out, switch modes or nudge the fixed slider and convert again from your saved original.

Pay attention to edges and shadows specifically: a band of solid black along one margin or a washed-out corner usually means the wrong mode for that lighting. Also sanity-check the downloaded file size against your goal — if you converted to shrink an archive, confirm the PNG is actually smaller than the source.

One caution for documents that carry weight. For anything official, financial, or legal — contracts, IDs, invoices, signed records — verify the black-and-white version against the original by eye at full zoom before you submit, file, or share it. A threshold can quietly erase a pale digit, a faint stamp, or a light signature, and an automated conversion is not a substitute for that human check.