Is there the possibility to hook in OCR for text blocks flattened into an image, maybe with some callback? That’s my biggest gripe with dealing with PDFs.
Yeah, sorry for confusion. When said Unicode, meant foreign text rather (just) the unescaped symbols, e.g. Greek. At one random Greek textbook[0], zpdf output is (extract | head -15):
Lol, but there's 100 competitors in the PDF text extraction space, some are multi million dollar industries: AWS textract, ABBY PDFreader, PDFBox, I think you may be underestimating the challenge here.
- Memory-mapped file I/O (no read syscalls)
- Zero-copy parsing where possible
- SIMD-accelerated string search for finding PDF structures
- Parallel extraction across pages using Zig's thread pool
- Streaming output (no intermediate allocations for extracted text)
What it handles:
- XRef tables and streams (PDF 1.5+)
- Incremental PDF updates (/Prev chain)
- FlateDecode, ASCII85, LZW, RunLength decompression
- Font encodings: WinAnsi, MacRoman, ToUnicode CMap
- CID fonts (Type0, Identity-H/V, UTF-16BE with surrogate pairs)
FWIW - mupdf is simply not fast. I've done lots of pdf indexing apps, and mupdf is by far the slowest and least able to open valid pdfs when it came to text extraction. It also takes tons of memory.
a better speed comparison would either be multi-process pdfium (since pdfium was forked from foxit before multi-thread support, you can't thread it), multi-threaded foxit, or something like syncfusion (which is quite fast and supports multiple threads). Or even single thread pdfium vs single thread your-code.
These were always the fastest/best options. I can (and do) achieve 41k pages/sec or better on these options.
The other thing it doesn't appear you mention is whether you handle putting the words in reading order (IE how they appear on the page), or only stream order (which varies in its relation to apperance order) .
If it's only stream order, sure, that's really fast to do. But also not anywhere near as helpful as reading order, which is what other text-extraction engines do.
Looking at the code, it looks like the code to do reading order exists, but is not what is being benchmarked or used by default?
If so, this is really comparing apples and oranges.
We're well beyond benefit of the doubt these days. If it looks like a duck... For me there wasn't any doubt, the author's first top comment here was evidence enough, then seeing the readme + random code + random commit message, it's all obvious LLM-speak to me.
I don't particularly care, though, and I'm more positive about LLMs than negative even if I don't (yet?) use them very much. I think it's hilarious that a few people asked for Python bindings and then bam, done, and one person is like "..wha?" Yes, LLMs can do that sort of grunt work now! How cool, if kind of pointless. Couldn't the cycles have just been spent on trying to make muPDF better? Though I see they're in C and AGPL, I suppose either is motivation enough to do a rewrite instead. (This is MIT Licensed though it's still unclear to me how 100% or even large-% vibe-coded code deserves any copyright protection, I think all such should generally be under the Unlicense/public domain.)
If the intent of "benefit of the doubt" is to reduce people having a freak out over anyone who dares use these tools, I get that.
You avoid an unnecessary copy. Normal read system call gets the data from disk hardware into the kernel page cache and then copies it into the buffer you provide in your process memory. With mmap, the page cache is mapped directly into your process memory, no copy.
All running processes share the mapped copy of the file.
There are a lot of downsides to mmap: you lose explicit error handling and fine-grained control of when exactly I/O happens. Consult the classic article on why sophisticated systems like DBMSs do not use mmap: https://db.cs.cmu.edu/mmap-cidr2022/
it allows the program to reference memory without having to manage it in the heap space. it would make the program faster in a memory managed language, otherwise it would reduce the memory footprint consumed by the program.
I don't fully understand the under the hood mechanics of mmap, but I can sense that you're trying to convey that mmap shouldn't be used a blanket optimization technique as there are tradeoffs in terms of page fault overheads (being at the mercy of OS page cache mechanics)
Tradeoffs such as "if an I/O error occurs, the program immediately segfaults." Also, I doubt you're I/O bound to the point where mmap noticeably better than read, but I guess it's fine for an experiment.
I think he's conveying that he doesn't know what he's talking about. buf[i] generates the same code regardless of whether mmap is being used. The first access to a page will cause a trap that loads the page into memory, but this is also true if the memory is read into.
Impressive performance gains! 5x faster than MuPDF is significant, especially for applications processing large volumes of PDFs. Zig's memory safety without garbage collection overhead makes it ideal for this kind of performance-critical work.
I'm curious about the trade-offs mentioned in the comments regarding Unicode handling. For document analysis pipelines (like extracting text from technical documentation or research papers), robust Unicode support is often critical.
Would be interesting to see benchmarks on different PDF types - academic papers with equations, scanned documents with OCR layers, and complex layouts with tables. Performance can vary wildly depending on the document structure.
very nice, it'd be good to see a feature comparison as when I use mupdf it's not really just about speed, but about the level of support of all kinds of obscure pdf features, and good level of accuracy of the built-in algorithms for things like handling two-column pages, identifying paragraphs, etc.
the licensing is a huge blocker for using mupdf in non-OSS tools, so it's very nice to see this is MIT
It seems that he didn't even test it before submitting though…
The author has created 30 new projects on github, in half a dozen different programming language, over the past month alone, and he also happen to have an LLM-generated blog. I think it's fair to say it's not “legitimately useful” except as a way for the author to fill his resume as he's looking for a job.
Not being slow - they compile straight to bytecode, they aren't interpreted, and have aggressive, opinionated optimizations baked in by default, so it's even faster than compiled c (under default conditions.)
Contrasted with python, which is interpreted, has a clunky runtime, minimal optimizations, and all sorts of choices that result in slow, redundant, and also slow, performance.
The price for performance is safety checks, redundancy, how badly wrong things can go, and so on.
A good compromise is luajit - you get some of the same aggressive optimizations, but in an interpreted language, with better-than-c performance but interpreted language convenience, access to low level things that can explode just as spectacularly as with zig or c, but also a beautiful language.
Zig is safer than C under default conditions, not faster. By default does a lot of illegal behavior safety checking, such as array and slice bounds checking, numeric overflow checking, and invalid union access checking. These features are disabled by certain (non default) build modes, or explicitly disabled at a per scope level.
It may be easier to write code that runs faster in Zig than in C under similar build optimization levels, because writing high performance C code looks a lot like writing idiomatic Zig code. The Zig standard library offers a lot of structures like hash maps, SIMD primitives, and allocators with different performance characteristics to better fit a given use-case. C application code often skips on these things simply because it is a lot more friction to do in C than in Zig.
>you have the time and energy to do stuff like all the bullet points listed
Don't disagree but in specific case, per the author, project was made via Claude Code. Although could as well be that Zig is better as LLM target. Noticed many new vibe projects decide to use Zig as target.
If it's really better than what we had before, what does it matter how it was made? It's literally hacked together with the tools of the day (LLMs) isn't that the very hacker ethos? Patching stuff together that works in a new and useful way.
5x speed improvements on pdf text extraction might be great for some applications I'm not aware of, I wouldn't just dismiss it out of hand because the author used $robot to write the code.
Presumably the thought to make the thing in the first place and decide what features to add and not add was more important than how the code is generated?
That's a very big if. The whole point is that what we had before was made slowly. This was made quickly. In itself it's not better but what it typically means is hours and hours of testing. Going through painful problems that highlight idiosyncrasies of the problem space. Things that are really weird and specific to whatever the tool is trying to address.
In such cases we can be expect that with very little time very few things were tested and tested properly (including a comment mentioned how tests were also generated). "We" the audience of potentially interested users have then to do that work (as plenty did commenting on that post).
IMHO what you bring forward is precisely that :
- can the new "solution" actually pass ALL the tests the previous one did? More?
This should be brought to the top and the actual compromises can then be understood, "we" can then decide if it's "better" for our context. In some cases faster with lossy output is actually better, in others absolutely not. The difference between the new and the old solutions isn't binary and have no visibility on that is what makes such a process nothing more than yet another showcase that LLMs can indeed produce "something" that is absolutely boring while consuming a TON of resources, including our own attention.
TL;DR: there should be test "harness" made by 3rd parties (or from well known software it is the closest too) that an LLM generated piece of code should pass before being actually compared.
[0]: https://repository.kallipos.gr/handle/11419/15087
ΑΛΕΞΑΝΔΡΟΣ ΤΡΙΑΝΤΑΦΥΛΛΙΔΗΣ Καθηγητής Τμήματος Βιολογίας, ΑΠΘ
~41K pages/sec peak throughput.
Key choices: memory-mapped I/O, SIMD string search, parallel page extraction, streaming output. Handles CID fonts, incremental updates, all common compression filters.
~5,000 lines, no dependencies, compiles in <2s.
Why it's fast:
What it handles:a better speed comparison would either be multi-process pdfium (since pdfium was forked from foxit before multi-thread support, you can't thread it), multi-threaded foxit, or something like syncfusion (which is quite fast and supports multiple threads). Or even single thread pdfium vs single thread your-code.
These were always the fastest/best options. I can (and do) achieve 41k pages/sec or better on these options.
The other thing it doesn't appear you mention is whether you handle putting the words in reading order (IE how they appear on the page), or only stream order (which varies in its relation to apperance order) .
If it's only stream order, sure, that's really fast to do. But also not anywhere near as helpful as reading order, which is what other text-extraction engines do.
Looking at the code, it looks like the code to do reading order exists, but is not what is being benchmarked or used by default?
If so, this is really comparing apples and oranges.
You didn't. Claude did. Like it did write this comment.
And you didn't even bother testing it before submitting.
From https://github.com/Lulzx/zpdf/blob/main/src/main.zig it looks like the help text cites an unimplemented "-j" option to enable multiple threads.
There is a "--parallel" option, but that is only implemented for the "bench" command.
I haven't tested without SIMD.
Are you using LLMs for parts of the coding?
What's your work flow when approaching a new project like this?
I can't talk about the code, but the readme and commit messages are most likely LLM-generated.
And when you take into account that the first commit happened just three hours ago, it feels like the entire project has been vibe coded.
2. Be not good at or a fan of git when committing
Not sure what the disconnect is.
Now if it were vibecoded, I wouldn't be surprised. But benefit of the doubt
I don't particularly care, though, and I'm more positive about LLMs than negative even if I don't (yet?) use them very much. I think it's hilarious that a few people asked for Python bindings and then bam, done, and one person is like "..wha?" Yes, LLMs can do that sort of grunt work now! How cool, if kind of pointless. Couldn't the cycles have just been spent on trying to make muPDF better? Though I see they're in C and AGPL, I suppose either is motivation enough to do a rewrite instead. (This is MIT Licensed though it's still unclear to me how 100% or even large-% vibe-coded code deserves any copyright protection, I think all such should generally be under the Unlicense/public domain.)
If the intent of "benefit of the doubt" is to reduce people having a freak out over anyone who dares use these tools, I get that.
I'll try my best to make it a really good one!
You avoid an unnecessary copy. Normal read system call gets the data from disk hardware into the kernel page cache and then copies it into the buffer you provide in your process memory. With mmap, the page cache is mapped directly into your process memory, no copy.
All running processes share the mapped copy of the file.
There are a lot of downsides to mmap: you lose explicit error handling and fine-grained control of when exactly I/O happens. Consult the classic article on why sophisticated systems like DBMSs do not use mmap: https://db.cs.cmu.edu/mmap-cidr2022/
I now wonder which use cases would mmap suit better - if any...
> All running processes share the mapped copy of the file.
So something like building linkers that deal with read only shared libraries "plugins" etc ..?
[1] https://github.com/pdf-association/pdf-corpora
I'm curious about the trade-offs mentioned in the comments regarding Unicode handling. For document analysis pipelines (like extracting text from technical documentation or research papers), robust Unicode support is often critical.
Would be interesting to see benchmarks on different PDF types - academic papers with equations, scanned documents with OCR layers, and complex layouts with tables. Performance can vary wildly depending on the document structure.
https://github.com/Lulzx/zpdf/blob/main/python/tests/test_zp...
the licensing is a huge blocker for using mupdf in non-OSS tools, so it's very nice to see this is MIT
python bindings would be good too
also, added python bindings.
as others have commented, I think while this is a nice portfolio piece, I would worry about its longevity as a vibe coded project
The author has created 30 new projects on github, in half a dozen different programming language, over the past month alone, and he also happen to have an LLM-generated blog. I think it's fair to say it's not “legitimately useful” except as a way for the author to fill his resume as he's looking for a job.
This kind of behavior is toxic.
Contrasted with python, which is interpreted, has a clunky runtime, minimal optimizations, and all sorts of choices that result in slow, redundant, and also slow, performance.
The price for performance is safety checks, redundancy, how badly wrong things can go, and so on.
A good compromise is luajit - you get some of the same aggressive optimizations, but in an interpreted language, with better-than-c performance but interpreted language convenience, access to low level things that can explode just as spectacularly as with zig or c, but also a beautiful language.
It may be easier to write code that runs faster in Zig than in C under similar build optimization levels, because writing high performance C code looks a lot like writing idiomatic Zig code. The Zig standard library offers a lot of structures like hash maps, SIMD primitives, and allocators with different performance characteristics to better fit a given use-case. C application code often skips on these things simply because it is a lot more friction to do in C than in Zig.
machine code, not https://en.wikipedia.org/wiki/Bytecode
> The price for performance is safety checks
In Zig, non-ReleaseFast build modes have significant safety checks.
> luajit ... with better-than-c performance
No.
Don't disagree but in specific case, per the author, project was made via Claude Code. Although could as well be that Zig is better as LLM target. Noticed many new vibe projects decide to use Zig as target.
- commit message: LLM-generated.
- README: LLM-generated.
I'm not convinced that projects vibe coded over the evening deserve the HN front page…
Edit: and of course the author's blog is also full of AI slop…
2026 hasn't even started I already hate it.
If it's really better than what we had before, what does it matter how it was made? It's literally hacked together with the tools of the day (LLMs) isn't that the very hacker ethos? Patching stuff together that works in a new and useful way.
5x speed improvements on pdf text extraction might be great for some applications I'm not aware of, I wouldn't just dismiss it out of hand because the author used $robot to write the code.
Presumably the thought to make the thing in the first place and decide what features to add and not add was more important than how the code is generated?
That's a very big if. The whole point is that what we had before was made slowly. This was made quickly. In itself it's not better but what it typically means is hours and hours of testing. Going through painful problems that highlight idiosyncrasies of the problem space. Things that are really weird and specific to whatever the tool is trying to address.
In such cases we can be expect that with very little time very few things were tested and tested properly (including a comment mentioned how tests were also generated). "We" the audience of potentially interested users have then to do that work (as plenty did commenting on that post).
IMHO what you bring forward is precisely that :
- can the new "solution" actually pass ALL the tests the previous one did? More?
This should be brought to the top and the actual compromises can then be understood, "we" can then decide if it's "better" for our context. In some cases faster with lossy output is actually better, in others absolutely not. The difference between the new and the old solutions isn't binary and have no visibility on that is what makes such a process nothing more than yet another showcase that LLMs can indeed produce "something" that is absolutely boring while consuming a TON of resources, including our own attention.
TL;DR: there should be test "harness" made by 3rd parties (or from well known software it is the closest too) that an LLM generated piece of code should pass before being actually compared.
fpdf
jpdf
cpdf
cpppdf
bfpdf
ppdf
...
opdf