Crop marks. Bleed.
Padding. Resizing.
AI edits.

Two PDFs. A press-ready one for the printer, with every page upscaled to your target DPI, the bleed carried past the trim and crop marks in the margin. And a low-resolution proof PDF with PROOF across every page, small enough to email the client for sign-off and hard to mistake for the press file. A twelve-page job comes back as one twelve-page PDF, in the original order.

PDF, Illustrator .ai, TIFF, JPEG and PNG. Multi-page sources are split into one page each, and every page is cropped, padded and edited on its own before they come back as a single PDF.

A0 through A7, US Letter, Legal and Tabloid, postcards at 4x6, 5x7 and 6x9, the 99x210mm DL flyer, business cards at 90x55mm, 89x51mm and a 70x28mm mini card, square cards and square formats, an 8x10 inch photo print, and 18x24 and 24x36 inch posters. If yours isn't in the list, type the millimetres. Portrait or landscape either way.

You set the bleed in whole millimetres and it mirrors the artwork outward past the trim on all four edges. It reflects what the designer already drew, so it can't invent artwork that was never there. On texture, sky and flat colour you won't see it. On type or a logo near the trim you will, as a reversed ghost, so look at the proof before you commit to a full bleed on a busy edge.

Eight strokes, one pair at each corner, on every page. The length, the offsets, the weight and the colour are fixed in the code and there's no setting for them. A TrimBox goes into the PDF. There's no BleedBox or ArtBox, and no registration marks, colour bars or slug information. If your workflow needs more than corner marks, it has to come from somewhere else.

Pad it. That extends the canvas before you crop, so you have something to frame against instead of cutting into the design. The value is a percentage of the image's own dimension, from 1 to 100. Do one page or tick the box to do all of them. Applying it twice stacks a second bar on the first instead of replacing it, and every step is kept as a revision, so you can step back ten of them, redo, or restore a version you had abandoned.

Yes, from the crop screen. The crop box re-locks to the new ratio and you re-frame the page. No re-upload, and the file you sent in is never modified.

It gets you further than plain resampling does, but nothing puts back detail the camera never captured. Photographs under about two megapixels go through Real-ESRGAN, which invents plausible detail to fill the gap. That is fine on a landscape and risky on a product shot or on legal copy, where what the model invents becomes a defect you own. Line art and logos take a Lanczos path instead, because a model softens the edges you want kept hard. The proof PDF exists so you and the client agree on the result before it goes to press.

What's printed on the artwork: phone numbers, email addresses, names, taglines, background colours, the position or size of a logo. The suggestions it offers are content changes only, so it won't propose fixes to bleed, trim or crop marks. Those are the pipeline's job. Applying an edit does replace the whole page image at whatever size the model produced, which clears the crop and can change the resolution, which is why the app reports the DPI you ended up with. One page at a time, while the job is at the crop stage.

Usually 60 to 180 seconds. What comes back is a proposal and nothing lands on the page until you hit Apply, so you can discard it or refine it instead. The image from before the edit is kept, so one click puts it back. Because the model picks its own output size, the app measures what came back against your trim width and tells you the DPI you actually ended up with.

No. The rest of the pipeline works exactly the same without it. When it is available, attempts count against your organisation's monthly allowance of 200 by default. A generation that fails still counts, because it cost the same API call as a good one, but discarding an attempt deletes it and gives you that one back.

The pipeline works in RGB and hands back an RGB PDF, so if your printer needs a specific CMYK profile, that conversion is theirs to do. What Print Prepper does is read the source before rasterising and tell you what it found: spot colours, CMYK, intent. The loss shows up on the job instead of surprising you at the proof. Applying an AI edit discards those spot inks too, and the app says so before you commit.

Yes. Signing up creates an organisation, and you invite the rest of the shop into it by email. Everyone in the organisation can run jobs. Roles cover the account itself, so an admin can invite people, change roles and remove someone the day they leave. Authenticator-app two-factor is built in, with backup codes, and an owner can reset it for someone who has lost their phone.