I've been using gemma4 31b for document summarization and find it pretty good for that. (I use it via goose as the harness and give it a suite of PDF CLI tools like pdftotext.)
September 20, 2026 at 4:31 PM
I've been using gemma4 31b for document summarization and find it pretty good for that. (I use it via goose as the harness and give it a suite of PDF CLI tools like pdftotext.)
Couldn’t I also just… convert the text from hex to ascii and read it myself?
pdftotext -layout this_dumbass_file.pdf | xxr -r -p
should be pretty close?
pdftotext -layout this_dumbass_file.pdf | xxr -r -p
should be pretty close?
June 3, 2026 at 9:03 PM
Couldn’t I also just… convert the text from hex to ascii and read it myself?
pdftotext -layout this_dumbass_file.pdf | xxr -r -p
should be pretty close?
pdftotext -layout this_dumbass_file.pdf | xxr -r -p
should be pretty close?
One of the things I greatly lament is not learning PowerShell scripting when I started this road decades ago!
pdftotext is freaking awesome!!!!! Just saved hours a week breaking apart PDF to log ingesting. pdftotext takes minutes to get PDF logs ready!!!!!!!!
pdftotext is freaking awesome!!!!! Just saved hours a week breaking apart PDF to log ingesting. pdftotext takes minutes to get PDF logs ready!!!!!!!!
August 18, 2026 at 3:04 PM
One of the things I greatly lament is not learning PowerShell scripting when I started this road decades ago!
pdftotext is freaking awesome!!!!! Just saved hours a week breaking apart PDF to log ingesting. pdftotext takes minutes to get PDF logs ready!!!!!!!!
pdftotext is freaking awesome!!!!! Just saved hours a week breaking apart PDF to log ingesting. pdftotext takes minutes to get PDF logs ready!!!!!!!!
Time to start analysing these submissions.
I noticed that all of the form submissions had an "easyPDF" footer which includes a date. This text is stored [as text] in the PDF file, which allows the Linux pdftotext program to extract it, and easily parse it.
I noticed that all of the form submissions had an "easyPDF" footer which includes a date. This text is stored [as text] in the PDF file, which allows the Linux pdftotext program to extract it, and easily parse it.
April 2, 2025 at 8:04 PM
Time to start analysing these submissions.
I noticed that all of the form submissions had an "easyPDF" footer which includes a date. This text is stored [as text] in the PDF file, which allows the Linux pdftotext program to extract it, and easily parse it.
I noticed that all of the form submissions had an "easyPDF" footer which includes a date. This text is stored [as text] in the PDF file, which allows the Linux pdftotext program to extract it, and easily parse it.
PAC (PDF Accessibility Checker) is worth a look too, free and gives you a full tag tree instead of eyeballing it in a browser. poppler's pdftotext -layout is handy for a quick sanity check on reading order.
September 25, 2026 at 1:36 PM
PAC (PDF Accessibility Checker) is worth a look too, free and gives you a full tag tree instead of eyeballing it in a browser. poppler's pdftotext -layout is handy for a quick sanity check on reading order.
Und mit pdftotext wirklich identisch, obwohl Fettschrift meistens verschränkt doppelt ist wie bei ZZweitstimweitstimmendeckmendeckungsvungsvererfahrfahrenen.
July 30, 2024 at 4:28 PM
Und mit pdftotext wirklich identisch, obwohl Fettschrift meistens verschränkt doppelt ist wie bei ZZweitstimweitstimmendeckmendeckungsvungsvererfahrfahrenen.
Nerd thread on this:
The data is a pdf, in a table, but of course it's not a real table. Can't just copy stuff out nicely.
Fell into the same trap as when I mapped traffic calming requests thinking I could use pdftotext and cleverness to extract.
10 minutes with emacs rectangle mode got it done.
The data is a pdf, in a table, but of course it's not a real table. Can't just copy stuff out nicely.
Fell into the same trap as when I mapped traffic calming requests thinking I could use pdftotext and cleverness to extract.
10 minutes with emacs rectangle mode got it done.
New map from @danp.net!
I know it won’t surprise Halisky that not all intersections in Halifax have automatic pedestrian signals. But did you know there are a lot that do have automatic signals during the day, or are supposed to? And there’s no easy way to know if an intersection is on the list?
1/4
I know it won’t surprise Halisky that not all intersections in Halifax have automatic pedestrian signals. But did you know there are a lot that do have automatic signals during the day, or are supposed to? And there’s no easy way to know if an intersection is on the list?
1/4
Halifax Pedestrian Automatic Signal Locations
danp.net
February 16, 2025 at 5:17 PM
Nerd thread on this:
The data is a pdf, in a table, but of course it's not a real table. Can't just copy stuff out nicely.
Fell into the same trap as when I mapped traffic calming requests thinking I could use pdftotext and cleverness to extract.
10 minutes with emacs rectangle mode got it done.
The data is a pdf, in a table, but of course it's not a real table. Can't just copy stuff out nicely.
Fell into the same trap as when I mapped traffic calming requests thinking I could use pdftotext and cleverness to extract.
10 minutes with emacs rectangle mode got it done.
Good lord this tender site is awful. They put password protection on copying text out of the pdfs. pdftotext does not care about that.
February 19, 2025 at 8:14 PM
Good lord this tender site is awful. They put password protection on copying text out of the pdfs. pdftotext does not care about that.
@deboranis.bsky.social
Uma dica pra tirar distrações de formatações e transformar textos em pdf em txt:
pdftotext -layout input.pdf output.txt
www.linuxuprising.com/2019/05/how-...
Uma dica pra tirar distrações de formatações e transformar textos em pdf em txt:
pdftotext -layout input.pdf output.txt
www.linuxuprising.com/2019/05/how-...
How To Convert PDF To Text On Linux (GUI And Command Line)
This article presents 2 tools for converting PDF documents to editable text on Linux, using a graphical tool (Calibre) and a command line tool (pdftotext).
www.linuxuprising.com
July 15, 2025 at 2:16 PM
@deboranis.bsky.social
Uma dica pra tirar distrações de formatações e transformar textos em pdf em txt:
pdftotext -layout input.pdf output.txt
www.linuxuprising.com/2019/05/how-...
Uma dica pra tirar distrações de formatações e transformar textos em pdf em txt:
pdftotext -layout input.pdf output.txt
www.linuxuprising.com/2019/05/how-...
Así que naturalmente hice lo que cualquier persona haría y tomé las 5000 palabras más importantes para un estudiante de Inglés según Oxford y las procesé.
![cat <(pdftotext -layout American_Oxford_5000.pdf - | \tail +5) \
<(pdftotext -layout American_Oxford_3000.pdf - | tail +4) \
| sed 's/[^[:print:]]//g' | grep '^[a-zA-Z]' | sed 's/ \{3,\}/\n/g' \
| awk '{print $1}' | tr -d '[:digit:]' | sort -u \
| awk '{print length}' | sort | uniq -c | sort -nrk2 | sed 's/^[ ]*//g' \
| uplot bar --fmt yx -t "Number of words per character count"](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:gbq7kpqqmld2lnau5vp737yo/bafkreiaqa7uf3ggl6nkk4mjpmwga5o3ylm6tiaqbvvwbgmxh2c6fiuegza)
December 2, 2024 at 3:29 PM
Así que naturalmente hice lo que cualquier persona haría y tomé las 5000 palabras más importantes para un estudiante de Inglés según Oxford y las procesé.
dilemma... I respect your opinions on applied LLM questions... but I also respect Claude's opinions, and he always tries pdftotext first
May 30, 2026 at 3:38 AM
dilemma... I respect your opinions on applied LLM questions... but I also respect Claude's opinions, and he always tries pdftotext first
My suggestions:
1. Tabula (tabula.technology) for table extraction
2. Stripping text out while preserving formatting using pdftotext Unix utility
3. If you have Acrobat full version, combining files into smaller number and using export to Excel feature
1. Tabula (tabula.technology) for table extraction
2. Stripping text out while preserving formatting using pdftotext Unix utility
3. If you have Acrobat full version, combining files into smaller number and using export to Excel feature
September 9, 2023 at 2:20 PM
My suggestions:
1. Tabula (tabula.technology) for table extraction
2. Stripping text out while preserving formatting using pdftotext Unix utility
3. If you have Acrobat full version, combining files into smaller number and using export to Excel feature
1. Tabula (tabula.technology) for table extraction
2. Stripping text out while preserving formatting using pdftotext Unix utility
3. If you have Acrobat full version, combining files into smaller number and using export to Excel feature
New Module
image2text 🚀
Text from images / PDFs? Whether you're parsing lab reports, invoices, or scanned documents. image2text is a modular OCR pipeline
Clarity|Reproducibility|Control
📂 github.com/eddsosa/image2text
#Python #OCR #Tesseract #OpenSource #ScientificSoftware #image2text #PDFtoText
image2text 🚀
Text from images / PDFs? Whether you're parsing lab reports, invoices, or scanned documents. image2text is a modular OCR pipeline
Clarity|Reproducibility|Control
📂 github.com/eddsosa/image2text
#Python #OCR #Tesseract #OpenSource #ScientificSoftware #image2text #PDFtoText
GitHub - eddsosa/image2text
Contribute to eddsosa/image2text development by creating an account on GitHub.
github.com
November 7, 2025 at 1:16 AM
New Module
image2text 🚀
Text from images / PDFs? Whether you're parsing lab reports, invoices, or scanned documents. image2text is a modular OCR pipeline
Clarity|Reproducibility|Control
📂 github.com/eddsosa/image2text
#Python #OCR #Tesseract #OpenSource #ScientificSoftware #image2text #PDFtoText
image2text 🚀
Text from images / PDFs? Whether you're parsing lab reports, invoices, or scanned documents. image2text is a modular OCR pipeline
Clarity|Reproducibility|Control
📂 github.com/eddsosa/image2text
#Python #OCR #Tesseract #OpenSource #ScientificSoftware #image2text #PDFtoText
Linuxilla pdftotext ja vähän skripteillä automaatio monille tiedostoille.
March 20, 2026 at 11:25 AM
Linuxilla pdftotext ja vähän skripteillä automaatio monille tiedostoille.
antes/depois. pdftotext, some regex, two scripts written by gpt4, and a loop with a simple "clean up this text" prompt for chunks up to half the token limit to gpt3.5. total cost 50 cents and an evening, and all the code is reusable
April 14, 2023 at 1:13 PM
antes/depois. pdftotext, some regex, two scripts written by gpt4, and a loop with a simple "clean up this text" prompt for chunks up to half the token limit to gpt3.5. total cost 50 cents and an evening, and all the code is reusable
The Poppler package has some nice tools for getting text out of PDFs. Check out `pdftotext`, especially with the `-layout` option, and `pdftohtml`. I use them all the time for extracting things out of PDFs (such as invoice data).
pdftotext(1)
pdftotext − Portable Document Format (PDF) to text converter (version 4.05)
www.xpdfreader.com
January 17, 2025 at 1:36 PM
The Poppler package has some nice tools for getting text out of PDFs. Check out `pdftotext`, especially with the `-layout` option, and `pdftohtml`. I use them all the time for extracting things out of PDFs (such as invoice data).
aye! I tried the same Maine article creation on STORM (I had to make it more formal/data-driven). pdftotext + pdfx (for links) did a decent job on it too. rud.is/dl/maine-key/
rud.is
June 2, 2024 at 11:32 AM
aye! I tried the same Maine article creation on STORM (I had to make it more formal/data-driven). pdftotext + pdfx (for links) did a decent job on it too. rud.is/dl/maine-key/
Let's say you give it the rights to execute pdftotext and read the output, so that it can access new types of documents. Seems harmless right? What if there's a hole in pdftotext? Can someone craft a prompt and a pdf that exploits the hole?
February 8, 2026 at 9:51 AM
Let's say you give it the rights to execute pdftotext and read the output, so that it can access new types of documents. Seems harmless right? What if there's a hole in pdftotext? Can someone craft a prompt and a pdf that exploits the hole?
what the fck
is it so they dont even have to call pdftotext before feeding into an ai ats?
is it so they dont even have to call pdftotext before feeding into an ai ats?
June 15, 2026 at 9:41 PM
what the fck
is it so they dont even have to call pdftotext before feeding into an ai ats?
is it so they dont even have to call pdftotext before feeding into an ai ats?
$ pdftotext -layout fec-trump-2024-C00828541.pdf - | grep "P2024" | sed -e 's/ \ */|/g' > fec.psv
November 18, 2023 at 4:54 PM
$ pdftotext -layout fec-trump-2024-C00828541.pdf - | grep "P2024" | sed -e 's/ \ */|/g' > fec.psv
But then you're missing some functionalities. For example you'd like your agent to convert files into a readable format, which requires executing some commands (pdftotext or similar).
February 8, 2026 at 9:18 AM
But then you're missing some functionalities. For example you'd like your agent to convert files into a readable format, which requires executing some commands (pdftotext or similar).
If you read the post, it says that I used pdftotext (these PDFs didn’t need OCR) and used an LLM to produce structured JSON from unstructured text.
March 1, 2025 at 6:38 PM
If you read the post, it says that I used pdftotext (these PDFs didn’t need OCR) and used an LLM to produce structured JSON from unstructured text.
GUI with a Processing Dashboard, Results viewer, integrated PDF viewer with customization and ease of access.You can select an entire folder or multiple files at once.
github.com/KingBarker/u...
overbits.herokuapp.com/pdftotext/
github.com/KingBarker/u...
overbits.herokuapp.com/pdftotext/
GitHub - KingBarker/unredactGUI: A simple tool for reading poorly-redacted documents and attempt to reproduce their original form
A simple tool for reading poorly-redacted documents and attempt to reproduce their original form - KingBarker/unredactGUI
github.com
December 23, 2025 at 8:07 PM
GUI with a Processing Dashboard, Results viewer, integrated PDF viewer with customization and ease of access.You can select an entire folder or multiple files at once.
github.com/KingBarker/u...
overbits.herokuapp.com/pdftotext/
github.com/KingBarker/u...
overbits.herokuapp.com/pdftotext/
800pdf with 300 pages? this is an easy one: ls *.pdf | awk '{print "pdftotext -layout " $1 " -"}' | bash | egrep "emailadress" :-)
November 30, 2023 at 1:36 PM
800pdf with 300 pages? this is an easy one: ls *.pdf | awk '{print "pdftotext -layout " $1 " -"}' | bash | egrep "emailadress" :-)
These systems typically rely on backend text extraction libraries, most notably pdftotext (part of the Poppler open-source suite) or Apache Tika.
September 17, 2026 at 6:43 AM
These systems typically rely on backend text extraction libraries, most notably pdftotext (part of the Poppler open-source suite) or Apache Tika.