All activity
_stsaoleft a comment
After struggling to extract data from 300-page financial reports, I spent three months testing a dozen PDF solutions before building my own. While most tools are surprisingly accurate on the first pass, the real pain begins during the review process. I was tired of editing raw Markdown, constantly toggling between programs, and working without version control. That’s why I built OkraPDF....
OkraPDFFull-fidelity PDF Workbench
_stsaostarted a discussion
PDF Data Workbench - Early product feedback appreciated
Hi PH, I’d appreciate some early feedback on a PDF data tool I’ve spent the last three months building. My main frustration with existing solutions wasn't necessarily their initial accuracy, but rather the total lack of feedback loops. Specifically, I found that most tools I have used: Prevented quick fixes: You couldn't correct extraction or markdown errors directly alongside the source...
Inspired by https://labs.google.com/pomelli, I finally got to use my domain https://styleguide.fyi/ and run Claude Agent SDK in an Actor to make style guides for ~10 cents for tokens per run.
A page takes about 5 mins of active run and likely longer if there is a queue. The email input currently just saves to a Resend account that I'll manually trigger in this Alpha. If anyone uses it, there's no email campaign.
This is free for now and I hope you give it a try! Thanks.
styleguide.fyi alphastyle guide from your favorite brands