Due to pdf popularity there is a lot of demand for pdf processing tools. And the format is so complex that there are many nontrivial and creative ways to do pdf processing. That's why these "Hello World" projects usually make Top 5 on HN, and one of the upvotes is usually from me.
Your commands to process PDF with Ghostscript are lossy (they lose lots of metadata and in minor ways they also change how the PDF renders), and they produce very large PDF files.
During my college days, I used iLovePDF a lot, so I wanted to build an alternative to it. It’s not just about PDFs - I also have work in progress around image processing and related tools and Chrome Extetion as well
Sure, if they're tested well enough that there are no obvious UX issues (which is usually not the case)
It's just that there's zero effort put into them so they don't really offer anything of value. If you write a todo list-tier app, it would be completely useless to most people, but it's a learning project for you. If you vibecoded a todo list-tier app, it's completely useless to most people including yourself.
So if a platform is vibe-coded, it suddenly has no value? When the Spotify founder vibe-codes an app, it’s praised—but when an open-source contributor like me does it, it’s seen as a bad thing? That doesn’t seem fair
Was this done heavily LLM assisted? Especially the PDF Edit tools have user-interaction quirks and bugs that a human developer would catch immediately during the regular manual testing when developing.
I'd suggest you at least try and mitigate that by having the LLM do extensive e2e testing if you aren't interested in using your own product.
Yeah, and I have the feeling it is not tested at all.
It offers Word -> PDF conversion. Just for interest I tried it and it doesn't even get the simplest page right. It puts the filename into an header. The test page had 4 images, one svg, one pdf (from svg), and another variation of the first 2. The generated PDF only contains 2 of those images with wrong sizes. The later two are missing. So it's basically completely useless.
The free of charge LibreOffice gives much better results with its own caveats.
I don't even care about that. My suggestion to him was earnest.
I don't have a problem with LLMs. Just with how people use them. I just don't like "slop". I see the same user-interaction problems every time.
I just don't want people to litter their heavily polished immaculately styled products that have so clearly bad user-interaction design. E2e testing and closing the loop on LLMs does seem to help here.
Though I really would prefer people click around their own product for at least 5 minutes.
It matters to me. Depending on the ratio there is a line between 'LLM assisted' and 'LLM derived'. There are enough samples of open source code around this theme out there that this could be one of either and the goal to commercialize it is a messy one if the provenance of the code isn't clear. It would be great to see this sort of thing litigated so that there is at least some clarity rather than just a moral stance.
I wonder why everything now is written in web frameworks. Meanwhile I am currently using macos which has a magnificent PDF tool called... Preview. It allows annotate, merge, realign pages, insert one page from another document or even a JPEG-scan, etc.
However, before the courtesy of my company giving me a macos-enabled gear - I had to cope with PDFs using multiple apps on Windows and Linux. Recently I got there again and found out that PDF support is really weak in Linux, and the formerly award winning Acrobat Reader now looks slow and poor, trying to steal my data and occupy as much space as possible. Also Acrobat Reader reference browser for linux is killed now.
Hence, the question. If everyone is using PDF, why there are no good, fast native tools? and... why are we even staying with PDF?
How does it fare with PDFs consisting entirely of images? Any PDF tool was struggling with compressing a passport scan (made with iPhone so might've contributed somehow, knowing Apple and PDFs) I had to cut down in size. Ended up using ImageMagick cause any Ghostscript based tool couldn't get it below 7 MBs from the original 28MB which, although, pretty good, was still too high and I could tell there was still plenty of detail that could be discarded without losing the eligibility of the document. I had to compress it with ImageMagick at the end, cut it down from 28MB to 3MB.
Also does Adobe have some kind of patent/copyright on PDF forms? I don't think I saw any free tools that can edit fillable fields / tables in PDFs. I don't see any mention of forms in the Suite section of your app either. Is it just stupidly difficult / annoying to implement ?
Image-only PDFs (scans):
These are the hardest case. If a PDF is basically high-res images (like iPhone scans), browser-based tools have limits compared to ImageMagick, which has much finer control over resampling and JPEG compression. Ghostscript-style pipelines help, but ImageMagick often wins if you’re willing to discard more detail. Improving this is on the roadmap, but it’s genuinely tough in-browser.
PDF forms:
Adobe doesn’t own forms, but editable PDF forms are extremely complex and poorly standardized. Many free tools avoid true form editing because it’s easy to break files. That’s why I haven’t enabled it yet—possible, just time-consuming and error-prone.
For my hard sci-fi novel, I wanted people to give me feedback by annotating the PDF directly. Since I didn't know what local PDF editors they had available, I decided to vibe-code a web-based PDF annotation editor using PDF.js. (Yes, malicious users could have a field day by guessing the URLs.) It's pretty rough:
Basically, you drop a PDF onto your own web server. The web server serves up PDFs via PDF.js on the client. When the user highlights text to annotate it, the date, time, and text of all annotations in the document are pushed back to the server. As the author, when I reload the same PDF URL, I can add, review, modify, navigate through, or summarize the annotations just like a reader. Here's a screenshot with a funny comment one of my beta readers made:
Me and my buddy run a small indie dev studio, and a while back we got frustrated with how most PDF scanner apps feel — clunky UX, subscriptions everywhere, ads, and in some cases your documents get uploaded who-knows-where (for example, incidents reported leaks by TechRadar and Fox News).
So we built our own PDF scanner & editor — lightweight, privacy-first, and (hopefully) not annoying to use. No ads, no subscriptions. Most features are free — a couple of advanced tools require a one-time unlock. All core features run 100% offline with on-device processing.
The main features are built for everyday workflows:
Scan documents — auto edge detect, live corner adjust, batch multi-page
Fill and sign forms — reusable signatures, flatten for secure sharing
OCR text recognition — preserves layout, searchable PDFs or clean text export (supports 18 languages, e.g., English, Spanish, French, German, Italian, Portuguese, etc.)
Edit OCR-detected text — adjust or fix recognised text
Page tools — reorder, rotate, duplicate, delete, extract pages
Annotations and highlights — comments, text notes, custom watermarks
Folder organization — custom folders, drag-and-drop move/rename
Everything runs locally — no accounts, no tracking, no upload processing.
You can download an AI model to your device (one-time download — it stays cached), and then:
- ask questions about a document
- summarise sections or chapters
- extract key points or data
- turn long documents into quick notes
- After the model is installed, all Chat PDF processing happens fully offline on your device.
The app is free to download, and most features are free (scanning, OCR, signatures, annotations, editing, etc).
We wanted to keep the essential tools free, and only charge once for a few advanced features.
We also put together a YouTube playlist with short feature walkthroughs.
We’d really appreciate feedback — especially on the Chat PDF feature (usefulness, speed, UX, edge cases, things it should do better). If you try it and have suggestions, we’re actively improving the app based on user feedback.
After the incident with Tailwind CSS, I decided not to make this open source. Sponsorship has been zero since COVID, so it’s genuinely hard for open-source developers to sustain their work
I'm hoping one of these efforts will lead to local translation of PDFs. Anyone aware of one? Not local, but the best I've found is using Google Translate via camera/images.
I developed an aversion to "with love"-marketing. I've seen too many products come full circle from idealistic "ad-free-forever" "will-never-sell-your-data" "open-source-forever" "customer-first" student-times to selling out everything.
Just to be clear, I’ve been contributing to the open-source community since 2020, and I have no intention of misleading anyone. The use of the word ‘love’ isn’t about branding off another product-it’s simply a tool I personally needed, so I built it. If you’d like, you can also check my GitHub to see my work.
“Sorry about that. Some people jump to conclusions about things they don’t fully understand. I’m just trying to build something useful and contribute to the community, but reactions like this can make it difficult
It was always very hard to make money directly from open source. The Tailwind thing (incident??) was only notable in terms of its publicity. The fact that they thought they could make money from code the customer doesn't have to pay for is the incident perhaps?
There's a problem with i18n on the landing page, set my browser to German I see things like "home.alternative_title". Tbh I'm not sure such a site needs i18n at all, Claude was a bit overzealous there ;)
well, you just have to undercut Adobe Suite price by 50% and most people would buy your stuff:
Most people mainly use only PDf merge & split & add comments & signing - and pay for this a monthly Adobe subscription of 29,99 USD last time I checked.
Great work, thanks for sharing and congrats on the launch!
Very very small note - many clickable things on your site (the "explore" and "new task" buttons, the directory and blog links at the top, etc.) don't change the cursor to the css "cursor:pointer" (ie the clicky hand)
You might want to add `cursor-pointer` to your tailwind <button> elements
Yeah, I’ll do that. I have a Chrome extension that I’m planning to make paid, and I may also release a desktop version. I’m thinking of pricing it cheaply—around $2 for lifetime access
I’m doing everything locally, with no pricing on the extension for now. I do plan to make it paid later, but since all processing happens locally, your files never leave your device and remain completely safe.
also not just PDF the image processing also WIP will be done by next week
Might be better to provide a downloadable executable instead of asking the user to trust that the browser isn't doing what the browser was designed to do.
I disagree on that. I think that the main value of this kind of tools is "no installation required".
There are already free PDF editors that can be downloaded and installed once forever. What I used most is Libreoffice Draw: it imports a PDF, edit it as if it were a file in its own format, export as PDF again. It's not the only choice. Firefox has had a vanilla PDF editor since last year: download a PDF or drag one inside the browser window, edit it, save it. It's enough to add a PNG of my signature and fill out forms.
makes it difficult to verify that it runs locally. unobfuscated source is not available. important actions, like open a PDF, save edited PDF, will be stuck or error if you cut the internet after opening the site and only unstuck after you reenable internet. I get it's probably for speed
anyway, if you save the page in Chrome and serve it on a local server, it works even with internet disabled, so there's that.
Thanks for pointing this out. You’re right - some assets are currently loaded at runtime, which can cause actions to hang if the internet is cut mid-session. All PDF processing itself happens locally in the browser, and as you noticed, serving the page locally works fully offline. Improving offline behavior and making this easier to verify is on the roadmap
Good work! I do like that the tools are task centric and that means I don"t have to handle all sorts of things, I just quickly learn the three to four tools that I really need (as a person working in the real world). #pareto
Now, privacy, I love it! That "normal people" just store stuff in the cloud "it's on my phone", yeah ok, is one thing. It's another topic…
But since Gmail came out and was all the rage in nerd circles, I am wondering why the people who understand the tech the most, are so eager to hand over their data to Big Tech and some other very questionable entities.
Here's the thing in terms of money.
If your app does put my data into the cloud, I am not going to use it. At all. Ever.
If your app blesses me with a beautifully designed native GUI (or UI), instead of presenting itself in Electron slop to me, then I am already almost sold. Literally. I start to consider forking over some cash to you, dear developer of that beautifully designed, privacy respecting app.
I do use my browser to browse the web. I am not interested in a "secondary OS architecture" where I have to play sys admin for a range of "apps" aka plugins. Neither Chrome plugins (I don't use Chromium based stuff.) nor Wordpress plugins, nor Emacs "modes" are going to replace well done native programs.
You don't care enough about your project to provide a native program? Tells me, I shouldn't care either. Good buy.
For a high school student who survives on an allowance, paying $39 for an app may be a bit much, but not for an adult with an income.
Curation. A good maintained app store does all the "sys admin" stuff for me. No viruses, no weird installation procedures and so on.
This is why that works. Hassle-free. Locally-run, native app, means beauty and privacy.
I would pay for that. Happily. In fact, I have done so many times. The success of a plethora of developers with paid-for apps in the stores proves I am not the only one.
And, btw, this is the distribution/commerce model that RMS always favoured. I quote RMS:
> Since “free” refers to freedom, not to price, there is no contradiction between
selling copies and free software. In fact, the freedom to sell copies is crucial: collections of free software sold on CD-ROMs are important for the community, and selling them is an important way to raise funds for free software development. Therefore, a program that people are not free to include on these collections is not free software.
This is basically the app-store model.
And I would pay, for the above stated reasons and I would be inclined to gulp an even higher price if the package has the "OSS inside" sticker on it. For personal reasons, right?
Then there is one last thing. I don't want to have to create an account somewhere just to test-drive your app. Or to use it fully, later on.
Privacy means, I don't have to be online in order to use the software. The end.
Ah, cool idea. I’m currently integrating image processing features—crop, compress, and meme generation. Once that’s almost done, we can move on to integrating the workflow.
Show HN: PDF Quick – Free PDF tools with 100% client-side processing
https://news.ycombinator.com/item?id=46094734
Show HN: A privacy-first, client-side toolbox (PDF, Imgs, Dev) no server uploads
https://news.ycombinator.com/item?id=46018221
Show HN: FileZen – Client-side PDF and Video tools using WebAssembly
https://news.ycombinator.com/item?id=46339833
Show HN: JW Tool Box – Free, privacy-first web tools (PDF, Image, Converters)
https://news.ycombinator.com/item?id=46065448
Show HN: PDFClear – Browser-based PDF tools with local AI (WASM+Transformers.js)
https://news.ycombinator.com/item?id=46036944
Show HN: Free PDF tools that run in the browser
https://news.ycombinator.com/item?id=46315672
Show HN: Client-side file tools – PDF, images, crypto, all in-browser
https://news.ycombinator.com/item?id=46209627
They're all wrapping PDFlib and provide the same functionality.
Optimize PDF:
Merge PDF: And so on and so forth.It's just that there's zero effort put into them so they don't really offer anything of value. If you write a todo list-tier app, it would be completely useless to most people, but it's a learning project for you. If you vibecoded a todo list-tier app, it's completely useless to most people including yourself.
All with similar design, similar implementation, similar HN post. Literally AI slop.
Also:
Show HN: BentoPDF is a privacy first PDF Toolkit
https://news.ycombinator.com/item?id=46657892 (this one just yesterday)
Show HN: NoUploadTools – Free Tools that don't upload your files
https://news.ycombinator.com/item?id=46516400
Show HN: TechRex – client-side PDF editor (no upload, no watermark)
https://news.ycombinator.com/item?id=45611498
I'd suggest you at least try and mitigate that by having the LLM do extensive e2e testing if you aren't interested in using your own product.
It offers Word -> PDF conversion. Just for interest I tried it and it doesn't even get the simplest page right. It puts the filename into an header. The test page had 4 images, one svg, one pdf (from svg), and another variation of the first 2. The generated PDF only contains 2 of those images with wrong sizes. The later two are missing. So it's basically completely useless.
The free of charge LibreOffice gives much better results with its own caveats.
I just don't want people to litter their heavily polished immaculately styled products that have so clearly bad user-interaction design. E2e testing and closing the loop on LLMs does seem to help here.
Though I really would prefer people click around their own product for at least 5 minutes.
However, before the courtesy of my company giving me a macos-enabled gear - I had to cope with PDFs using multiple apps on Windows and Linux. Recently I got there again and found out that PDF support is really weak in Linux, and the formerly award winning Acrobat Reader now looks slow and poor, trying to steal my data and occupy as much space as possible. Also Acrobat Reader reference browser for linux is killed now.
Hence, the question. If everyone is using PDF, why there are no good, fast native tools? and... why are we even staying with PDF?
Also does Adobe have some kind of patent/copyright on PDF forms? I don't think I saw any free tools that can edit fillable fields / tables in PDFs. I don't see any mention of forms in the Suite section of your app either. Is it just stupidly difficult / annoying to implement ?
Image-only PDFs (scans): These are the hardest case. If a PDF is basically high-res images (like iPhone scans), browser-based tools have limits compared to ImageMagick, which has much finer control over resampling and JPEG compression. Ghostscript-style pipelines help, but ImageMagick often wins if you’re willing to discard more detail. Improving this is on the roadmap, but it’s genuinely tough in-browser.
PDF forms: Adobe doesn’t own forms, but editable PDF forms are extremely complex and poorly standardized. Many free tools avoid true form editing because it’s easy to break files. That’s why I haven’t enabled it yet—possible, just time-consuming and error-prone.
https://repo.autonoma.ca/?action=repo&repo=notanexus.git&vie...
Basically, you drop a PDF onto your own web server. The web server serves up PDFs via PDF.js on the client. When the user highlights text to annotate it, the date, time, and text of all annotations in the document are pushed back to the server. As the author, when I reload the same PDF URL, I can add, review, modify, navigate through, or summarize the annotations just like a reader. Here's a screenshot with a funny comment one of my beta readers made:
https://i.ibb.co/5gZMJ0qc/annotations.png
Beta readers wanted, see profile for contact!
So we built our own PDF scanner & editor — lightweight, privacy-first, and (hopefully) not annoying to use. No ads, no subscriptions. Most features are free — a couple of advanced tools require a one-time unlock. All core features run 100% offline with on-device processing.
The main features are built for everyday workflows:
Scan documents — auto edge detect, live corner adjust, batch multi-page Fill and sign forms — reusable signatures, flatten for secure sharing OCR text recognition — preserves layout, searchable PDFs or clean text export (supports 18 languages, e.g., English, Spanish, French, German, Italian, Portuguese, etc.) Edit OCR-detected text — adjust or fix recognised text Page tools — reorder, rotate, duplicate, delete, extract pages Annotations and highlights — comments, text notes, custom watermarks Folder organization — custom folders, drag-and-drop move/rename Everything runs locally — no accounts, no tracking, no upload processing. You can download an AI model to your device (one-time download — it stays cached), and then:
- ask questions about a document - summarise sections or chapters - extract key points or data - turn long documents into quick notes - After the model is installed, all Chat PDF processing happens fully offline on your device.
The app is free to download, and most features are free (scanning, OCR, signatures, annotations, editing, etc).
We wanted to keep the essential tools free, and only charge once for a few advanced features.
We also put together a YouTube playlist with short feature walkthroughs.
You can find the app here: https://apps.apple.com/ro/app/pdf-master-scan-edit-sign/id67...
We’d really appreciate feedback — especially on the Chat PDF feature (usefulness, speed, UX, edge cases, things it should do better). If you try it and have suggestions, we’re actively improving the app based on user feedback.
Buy-me-coffee / you can donate / payments in bitcoins accepted / pay as you use / etc.
But I am curious what could work so people wouldn’t be discouraged immediately?
Subscription (monthly/quarterly/annual) is annoying as well…
Adobe has started this wave, I remember it vividly.
How? Who?
Most of them are freemium, so they're balancing resources funded by subscriptions against the majority free user usage.
And is this local first (as it says on the website) or local only?
Very very small note - many clickable things on your site (the "explore" and "new task" buttons, the directory and blog links at the top, etc.) don't change the cursor to the css "cursor:pointer" (ie the clicky hand)
You might want to add `cursor-pointer` to your tailwind <button> elements
Built a client only webapp myself and offline usage is the main thing users ask about.
also not just PDF the image processing also WIP will be done by next week
There are already free PDF editors that can be downloaded and installed once forever. What I used most is Libreoffice Draw: it imports a PDF, edit it as if it were a file in its own format, export as PDF again. It's not the only choice. Firefox has had a vanilla PDF editor since last year: download a PDF or drag one inside the browser window, edit it, save it. It's enough to add a PNG of my signature and fill out forms.
Local-only web apps are great one-off projects, but extensions and native apps require much more maintenance.
Running an executable is a risk by default and the way it interacts with my network is way less transparent. I honestly prefer this in the browser.
anyway, if you save the page in Chrome and serve it on a local server, it works even with internet disabled, so there's that.
even if it might not stand before court it is enough for a lawyer to write you a letter that is not 100% baseless.
Now, privacy, I love it! That "normal people" just store stuff in the cloud "it's on my phone", yeah ok, is one thing. It's another topic…
But since Gmail came out and was all the rage in nerd circles, I am wondering why the people who understand the tech the most, are so eager to hand over their data to Big Tech and some other very questionable entities.
Here's the thing in terms of money.
If your app does put my data into the cloud, I am not going to use it. At all. Ever.
If your app blesses me with a beautifully designed native GUI (or UI), instead of presenting itself in Electron slop to me, then I am already almost sold. Literally. I start to consider forking over some cash to you, dear developer of that beautifully designed, privacy respecting app.
I do use my browser to browse the web. I am not interested in a "secondary OS architecture" where I have to play sys admin for a range of "apps" aka plugins. Neither Chrome plugins (I don't use Chromium based stuff.) nor Wordpress plugins, nor Emacs "modes" are going to replace well done native programs.
You don't care enough about your project to provide a native program? Tells me, I shouldn't care either. Good buy.
For a high school student who survives on an allowance, paying $39 for an app may be a bit much, but not for an adult with an income.
Curation. A good maintained app store does all the "sys admin" stuff for me. No viruses, no weird installation procedures and so on.
This is why that works. Hassle-free. Locally-run, native app, means beauty and privacy.
I would pay for that. Happily. In fact, I have done so many times. The success of a plethora of developers with paid-for apps in the stores proves I am not the only one.
And, btw, this is the distribution/commerce model that RMS always favoured. I quote RMS:
> Since “free” refers to freedom, not to price, there is no contradiction between selling copies and free software. In fact, the freedom to sell copies is crucial: collections of free software sold on CD-ROMs are important for the community, and selling them is an important way to raise funds for free software development. Therefore, a program that people are not free to include on these collections is not free software.
This is basically the app-store model.
And I would pay, for the above stated reasons and I would be inclined to gulp an even higher price if the package has the "OSS inside" sticker on it. For personal reasons, right?
Then there is one last thing. I don't want to have to create an account somewhere just to test-drive your app. Or to use it fully, later on.
Privacy means, I don't have to be online in order to use the software. The end.
Can we add workflows to this?
First merge all files then depending on output size compress to fit the size and other requirements?
Or take out page 35, then compress rest
Or extract page 2,5 and merge them and give me output withoit compress