Initializing... drag & drop files here
Supports: PS
.ps is a PostScript file — a program in Adobe's page-description language that paints a page when a PostScript interpreter runs it. WebP is the opposite kind of thing: one finished raster image, built for the web, roughly a quarter smaller than the JPEG or PNG equivalent. Getting from one to the other means rasterizing the page, and on this site that rasterizer is a PDF renderer, so the PostScript has to become a PDF first. A genuine .ps uploaded here does not come back as a WebP: the renderer looks for PDF structure (%PDF- and a cross-reference trailer), a PostScript stream (%!PS-Adobe-…) has neither, and the job ends in an error. Make the PDF yourself — one Ghostscript command — and PDF to WebP finishes the job with the exact controls described below.
.ps to PDF on Your Own Computer: ps2pdf input.ps output.pdf does it in one line. ps2pdf wraps Ghostscript's pdfwrite device, which is free for Windows, macOS and Linux; any "Print to PDF" or "Save as PDF" route from an application that can print the file works just as well.WebP is two codecs wearing one extension, and the toggle chooses between them. That choice matters more than the DPI for a page full of text, because lossy WebP is a VP8-derived transform that puts ringing around sharp black-on-white edges, while lossless WebP is an entropy coder that reproduces them exactly.
| Lossless? = No (preselected) | Lossless? = Yes | |
|---|---|---|
| Coding | Lossy VP8, 8-bit Y'CbCr 4:2:0 | Lossless, 8-bit ARGB |
| Text and line edges | Softened; artefacts around fine type at lower presets | Reproduced exactly |
| Typical size versus the alternative | 25–34% smaller than a comparable JPEG at equal SSIM | 26% smaller than the equivalent PNG |
| What the Quality Preset does | Sets the fidelity dial: Highest 100, Very High 95, High 90, Medium 80, Low 70, Very Low 60, Lowest 50 | Sets compression effort only — the pixels come out identical at every preset, and the file size shifts by a few percent |
| Pick it for | Pages dominated by photographs, scans or gradients | Diagrams, plots, forms and anything text-heavy |
The size figures come from Google's own WebP comparison, measured against PNG and JPEG at matched quality.
PostScript is Turing-complete. A .ps file does not describe a page so much as compute one, which is why rendering it needs a full interpreter — Ghostscript — rather than a parser. PDF took the same imaging model and threw away the programming language, leaving a static object graph that a parser can walk. Our page pipeline is built on the parser, so it reads PDF and not PostScript.
Two practical consequences follow. First, a real PostScript upload fails outright rather than producing a partial image, so changing the DPI or the Lossless toggle and retrying will not help. Second, if a file named .ps does convert here, it was PDF data all along — renderers identify files by their header bytes, not their extension, so a renamed PDF passes straight through. That is a useful test in itself: open the file in a PDF viewer, and if it opens, you already have the PDF and can skip step 1.
| Target | Compression | Alpha on this route | Best for |
|---|---|---|---|
| WebP (this page) | Lossy VP8 or lossless, your choice | Flattened onto the chosen colour | Web pages, small previews, anything bandwidth-sensitive |
| PDF to PNG | Deflate; truly lossless only at the Highest preset | Preserved when Image Transparency is left Unchanged | Screenshots of pages, diagrams, alpha work |
| PDF to JPG | JPEG, always lossy | None | Maximum compatibility with old software |
| PDF to TIFF | LZW, Deflate, PackBits, JPEG or none | Rarely used by readers | Print and archival hand-off |
| PDF to HEIC | HEVC, lossy | None | Apple-only workflows |
Because a PDF parser cannot read a PostScript program. The renderer expects the %PDF- header and a valid cross-reference trailer; .ps files begin %!PS-Adobe- and have neither, so the parse throws and the job reports an error with nothing partial written. Run ps2pdf input.ps output.pdf first, then upload the PDF to PDF to WebP — same renderer, same options, working result.
Lossless, in most cases. A rendered page is flat colour and hard edges, which is exactly what the lossless coder handles well and what lossy VP8 blurs — you will see haloing around small type at the lower presets. Lossy is the better choice when the page is mostly photographic, where it earns its 25–34% saving over JPEG without a visible cost.
Not to the image. In lossless mode the preset acts as a compression-effort dial: we encoded the same rendered page across the range and every output was pixel-identical, with only a few percent difference in file size between the extremes. In lossy mode the same preset is a genuine fidelity control, which is why the two modes need different habits.
Not on this route. WebP itself supports alpha in both lossy and lossless modes — Google puts the cost at about 22% extra bytes for lossless — but the option to preserve page transparency is exposed only for PNG output here, so the page is composited onto the Image Transparency colour. Choose PDF to PNG when you need the alpha channel to survive.
For an image that will be displayed at page-thumbnail size, 96 or 150 DPI is plenty and keeps the file small. For a preview a reader might zoom into, render at 300 DPI and then use Image resolution to bring it down to the display size — downscaling a high-resolution render preserves glyph shapes far better than rendering coarse in the first place. Reserve 600 and 1200 DPI for fine line art.
Each page becomes its own WebP. A four-page document gives four numbered files, downloadable individually or as a ZIP; there is no single WebP that holds a stack of pages, because WebP is a single-image format. The PDF you made in step 1 is already the multi-page container if you need one.
No. Rasterizing writes pixels and discards the text objects, so nothing in the output can be selected, searched or edited — the same is true of every image format on this list. Keep the intermediate PDF alongside the image if the text needs to stay live.
Uploads travel over an encrypted connection and are rendered on our servers; this is not an offline tool. Uploads and outputs are deleted automatically after a few hours, are never shared or made public, and no account is needed. There is no watermark on anything we produce and no sign-up between you and the download.