In short
This converts a .psd (or .psb) into aJPG of the flattened composite — the single merged image Photoshop stores next to your layers. It runs in your browser using a pure-TypeScript PSD engine: the file is read locally, the composite is rebuilt, and a JPEG comes back for download. JPG is the right target when you want a small, universally openable file and don't need transparency or pixel-exact fidelity. Nothing is uploaded, and you never need Photoshop.
How the conversion works
A PSD is a small 8BPS header (dimensions, colour mode, bit depth) followed by four length-prefixed sections: colour-mode data, image resources (which include a small embedded JPEG thumbnail), the layer and mask section, and finally the merged image data — the flattened composite. The engine reads that composite locally; for 8-bit Grayscale and 8-bit RGB documents it hands the pixels to a canvas, which re-encodes them to JPEG.
Two honest consequences follow from choosing JPG. First, JPG is lossy: the pixels are re-compressed at quality 0.92, so this is not a byte-exact copy — for a pixel-perfect result use the losslessPSD to PNG converter. Second, JPG has no alpha channel, so any transparency is flattened onto a white background before encoding. In exchange you get a much smaller file that opens in anything — email, a browser, any phone — without a Photoshop licence anywhere in sight.
What it can and can't do — honestly
What you get
- A high-quality JPEG (q0.92) of the flattened composite
- Transparency flattened onto white, so it opens cleanly anywhere
- 8-bit Grayscale and 8-bit RGB documents supported
- The embedded Photoshop thumbnail when that's the only preview left (labelled as such)
- A local read — every byte stays on your machine
What it won't do
- Keep transparency — JPG can't; use PNG if you need the alpha
- Give you a pixel-exact copy — JPG re-compresses; PNG is the lossless option
- Recover layers, masks, adjustment layers or effects — those aren't in the composite
- Render 16-bit / 32-bit, CMYK or Lab documents in-browser
- Invent pixels past a truncation — only the surviving rows come back
A note on the word "verified": it means the image we produced re-parses cleanly on our side, not that the source PSD is byte-for-byte perfect in Adobe's reader. And only the flattened composite is ever touched — the layer stack is never reconstructed, because once the layer section is damaged there's no deterministic way to rebuild it, and a guess dressed up as a recovery isn't honest.
If the PSD is damaged
If the composite can't be rebuilt — a truncated save from a crash or an interrupted Dropbox / Drive / Creative Cloud sync, or a corrupt header or section-length field — the tool won't fabricate a JPEG. It reports what it found and links the PSD repair page, which re-derives the section boundaries around the surviving bytes. Repair first, then convert the repaired file here.
Why "no upload" is verifiable, not a slogan
The PSD is read with the browser's file APIs and decoded on your machine; the page makes no request that carries your data. Don't take it on faith — open DevTools, select the Network tab, drop the file, and watch nothing leave. It's the same zero-upload principle as thezero-upload proof page, applied to conversion.