Ask HN: Whould you be interested in an PDF customization API? | Hacker News Reader

Ask HN: Whould you be interested in an PDF customization API?

1 pointtrez8 comments

We are building an API which allow to integrate some external data (POST request) to an existing design. That allow to integrate easily web2print feature in any website.

View on HN

I can just emit PS and use ps2pdf. PS is just text, so any language can do that. There's already an API - it's called Unix pipes!

never heard of that ;) Indeed, manual processing is an option if you are ok to add some code any time a user want a new PDF design. That's not a really scalable process if your number of PDFs is increasing. It's also not a viable option for less technical users

I'm sure it's computationally equivalent to what you're doing - it's a function from some input data and some template to a PDF file - so I don't follow your thesis that it would be less scalable (for some definition of scalable that you don't give).

You could even use some tech like Handlebars to do it with Node.js, so using your existing stack.

By scalable I mean accessible to a large number of user because that won't require too specific skills. My assumption is CSS is a more common skills than linux commands

How would this be different from the multiple products already in the market?

FPDFI, PDFLIB and Zend_PDF wkhtmltopdf flying-saucer

- No library to integrate (ie langage independant and you don't have to change your code) whereas actual product requires to use java, .net or python - easy customization with common technologies (CSS, JSON) whereas you have to use complex XML configuration with actual product - Per usage pricing instead of expensive license - Out-of-the-box widget to integrate in your website which adapt itself to the requested data

All theses library are about creating a PDF from scratch whereas our API should allow to customize an already existing PDF

Editing draw stream is creation. Once you parse the original the only real way to update the original stream is mark invalid and push in the updated stream.

If you are only looking to edit xmp/meta/comments then you could possibly use an edit or amend but would still need one of the pdf libraries to post process.

Reply on news.ycombinator.com