ChatGPT used to be able to read text from scanned PDF documents automatically. This was really helpful for my work and studying. I could just upload any PDF and it would extract the text even if it was just an image.
But now this feature doesn’t work anymore. When I upload a scanned PDF, ChatGPT doesn’t give me any response at all. No error message or anything. It just stays quiet.
I found out that this text recognition feature is now only available for Enterprise accounts. They removed it for regular paying subscribers without telling anyone. No announcement or warning about this change.
Now I have to use other tools to convert my PDFs first, or upload them as image files instead. This makes everything much more complicated and slow.
Has anyone else noticed this problem? It’s really annoying when a feature you depend on just disappears without warning. Is there any official explanation from OpenAI about why they did this?
Hit this same problem three weeks ago - totally blindsided me. Had a stack of academic papers to process and boom, nothing worked. OpenAI didn’t say a word about the change. No email, no dashboard alert, nothing.
I switched to Adobe Acrobat’s OCR, then fed that text to ChatGPT. Extra step, but the OCR’s actually pretty good. Google Drive works too if you upload PDFs there first.
What really bugs me is how they keep moving features behind higher paywalls without telling anyone. Hard to build reliable workflows when stuff changes overnight.
Yeah, this is exactly why automation beats vendor lock-in. Companies constantly change features and pricing - you shouldn’t rebuild your entire workflow each time.
I’ve been burned by services that suddenly drop features or spike prices. Solution? Build your own pipeline that you control.
Set up automated OCR processing with multiple providers. Scanned PDF comes in, gets processed through OCR engines, text extracted, then fed to whatever AI service you want. No more manual conversions or worrying about changes.
Best part - you can switch OCR services or AI providers without changing your process. ChatGPT drops another feature tomorrow? Your workflow keeps running.
Built something similar for work documents. Takes any scanned doc, processes automatically, outputs clean text for analysis. Haven’t worried about vendor changes since.
Latenode makes this automation simple to build and maintain. Check it out: https://latenode.com
OpenAI’s getting ridiculous. Same thing hit me two weeks back - thought it was just buggy but they quietly moved it to enterprise only. That’s like 10x the cost. I switched to Google Docs instead - upload your PDF and it auto-converts to text, then copy-paste into ChatGPT. Not perfect but works fine for most scanned docs.
Had this exact problem last month with research documents. Super frustrating - the feature worked great, then just stopped working with zero heads up in the interface or billing changes. I’ve been using a combo approach since then. For quick jobs, I screenshot the PDF pages and upload as images - ChatGPT still reads text from regular image files. For bigger documents, I run them through Microsoft OneNote’s free OCR first, then copy the extracted text. The screenshot workaround isn’t perfect since you’re stuck with upload limits and have to split multi-page docs, but it works for most stuff. OneNote’s OCR is actually pretty accurate, though you lose formatting. Really wish OpenAI would give us a heads up about these changes. Pulling features from paid plans without warning makes it hard to count on their platform for consistent work.