You can upload a PDF to ChatGPT using the Attachment function indicated by the “+” icon and selecting Add photos & files; the steps are nearly identical on both desktop and mobile devices. After uploading, ChatGPT will read the text in your document so you can ask questions about it.

Below, we will walk you through the upload steps, show you how to fix common issues, and provide tips to help you prepare your PDFs to get the best results with ChatGPT. We also cover what to ask ChatGPT once your file is loaded and how to protect your data after you upload it.

Key Takeaways

ChatGPT accepts files up to 512MB each on all plans, and text and document files, including PDFs, are also capped at 2 million tokens per file. Free accounts are limited to three uploads per day.
After logging in, you can start a new chat, click the Attachment icon, and select your PDF from your storage to upload it; the steps are nearly identical on desktop and mobile.
ChatGPT can’t read scanned (image-only) PDFs and only analyzes the text in PDFs that mix text and images.
Your uploaded files may be used to improve ChatGPT unless you turn off model training in Data Controls or use Temporary Chat, so think twice before uploading sensitive documents.

Can You Upload a PDF to ChatGPT?

Yes, you can upload documents to ChatGPT on the Free, Plus, Business, and Enterprise plans. Each file can be up to 512MB, and each custom GPT can hold up to 20 files over its lifetime.

You can upload up to 80 files every three hours (OpenAI may lower this limit during peak times), but if you are a Free user, you are limited to just three files per day.

The file-count limits for Plus and Business plans are based on Projects; Plus plan users are allowed to upload up to 20 files per project, while the maximum for the remaining plans is 40 files. These are file counts, not file sizes, so the 512MB per-file limit still applies. Check OpenAI’s help center for the latest numbers on your plan.

Aside from the PDF format, ChatGPT-supported file types also include images, spreadsheets, and presentations, with size limits as follows:

Images: up to 20MB per image
Spreadsheets: up to 50MB per spreadsheet
Presentations: up to 512MB per presentation

ChatGPT uses models that count tokens instead of words. A token is a piece of text; it can be a word, part of a word, or just punctuation. Most English words are one to two tokens each. Each text or document file, including PDFs, is limited to 2 million tokens (roughly 1.5 million words); spreadsheets are exempt from this cap.

In addition to the ChatGPT limits above, there is a 10GB usage cap for users and a 100GB usage cap for organizations.

How to Upload a PDF to ChatGPT: Desktop Version

Here is how to upload a PDF to ChatGPT on a desktop device:

Open ChatGPT and log in to your account.
Click New chat to start a conversation with the AI chatbot.
Click the Attachment (+) icon and select Add photos & files from the drop-down menu.
Select your PDF in the file window that opens and click Open. You can also drag and drop your PDF straight into the chat window. On plans that support it, you can attach files from Google Drive or Microsoft OneDrive in the same menu.
Wait a few seconds for ChatGPT to analyze the PDF.
Enter your prompt when the PDF loads fully and ask ChatGPT to do whatever you need with it.

How to Upload a PDF to ChatGPT: Mobile Version

Follow these steps to use ChatGPT with a PDF on mobile (menu labels may vary slightly between iPhone and Android):

Download the ChatGPT app and log in to your account.
Press the Attachment (+) icon and choose Add photos & files from the drop-down menu.
Select Choose files and upload your PDF.
Give the ChatGPT PDF reader a few seconds to process.
Ask the chatbot to extract any information you need from your PDF or perform any other activity required once it is loaded.

What to Ask ChatGPT After You Upload a PDF

Once your file has loaded, the quality of your prompt decides the quality of the answer. Here are prompts you can copy and adapt:

Summarize: “Summarize this PDF in five bullet points and list the three most important takeaways.”
Find key facts: “List every date, deadline, and dollar amount in this document, with page numbers.”
Explain a section: “Explain the section on [topic] in plain English and tell me which page it is on.”
Extract a table: “Turn the table on page [number] into a bulleted list.”
Review an agreement: “What obligations, risks, or fees does this contract mention? Quote the relevant lines and cite the pages.”
Study: “Create ten quiz questions with answers from this chapter.”
Translate: “Translate the executive summary into Spanish.”
Compare: “Compare the two PDFs I uploaded and list the main differences.”

ChatGPT can misread tables or invent details, so ask it to cite page numbers and double-check important figures against the original file. Numbering your pages (see the tips below) makes this much easier.

Why You Can’t Upload a PDF to ChatGPT: Troubleshooting Tips

You can’t upload a PDF to ChatGPT if you encounter the following common issues:

Problem	Likely cause	Quick fix
No upload option appears	Upload limit reached, or an app or browser glitch	Wait for the limit to reset or refresh the app (see #1)
The file won’t upload or the chat freezes	The file exceeds 512MB or 2 million tokens, or is very heavy	Compress or split the PDF (see #2)
ChatGPT can’t read the file	The PDF is a scan (image-only)	Convert the pages to images or run OCR (see #3)
Answers skip tables, images, or form fields	Complex formatting, embedded images, or a locked PDF	Paste the text, split the file, or remove restrictions (see #4)

#1. The Upload Button Isn’t Showing

The upload button might not show after clicking the Attachment icon if you have reached your upload limit (Free users get three files per day, and paid plans reset every three hours). In this case, you just need to wait for those limits to reset. If you haven’t hit a limit, update the ChatGPT app or refresh the page, log out and back in, try another browser, or disable extensions that may block uploads.

Upgrading to a higher-tiered plan is another option, as your usage cap will be expanded. However, this isn’t recommended for one-off PDF processing. Another option for one-off documents is to upload your document to pdf.net and use our Chat PDF tool instead.

#2. The PDF File Is Too Large

If the PDF is larger than 512MB or holds more than 2 million tokens of text, ChatGPT won’t be able to process it. Sometimes, even if your file hasn’t exceeded this size, the platform’s limited bandwidth during peak usage hours can cause freezes, lags, and crashes. A workaround is to compress your PDF in advance. You can also split it into smaller parts, as explained in the tips below.

You can do this in five steps with our PDF editor:

Open pdf.net, hover over More tools, and choose Compress from the options available.
Upload your document by dragging and dropping it into the Drop file here box or using the Choose file button.
Wait for the tool to process your PDF; it should take only several seconds.
Pick your preferred compression level.
Click Compress and download.

Besides resizing your document, you can also manually remove pages that you don’t need the AI to process, such as the cover page, table of contents page, or bibliography pages.

Besides resizing your document, you can also remove pages that you don’t need the AI to process, such as the cover page, table of contents page, or bibliography pages.

#3. The PDF Is Scanned or Not Readable

If your PDF is made up of scans or images, ChatGPT won’t be able to read it (OpenAI doesn’t support image-only PDFs). When this happens, you have two options:

Convert your PDF to image files. ChatGPT accepts images as a separate file type (up to 20MB each) and can read the text in them. You can do this with our PDF-to-JPG or PDF-to-PNG tools; both are browser-based, so there is no need for you to download any apps. After conversion, it is best to upload your images individually, as ChatGPT handles them better that way.
Use a separate OCR (Optical Character Recognition) PDF tool. It will turn your image-based PDF into a selectable, searchable one, which the AI chatbot can read. Afterwards, you can reupload the PDF to your chat/conversation. This route is better for multi-page scans because you keep a single document.

#4. ChatGPT Can’t Read Part of the PDF

Sometimes, despite correctly following the ChatGPT file upload guide above, the AI can’t read part of your PDF because it:

Doesn't just contain machine-readable text, but also other content stored as embedded elements. ChatGPT only analyzes the text in PDFs that mix text and images. A simple solution is to manually paste the problematic parts into ChatGPT; you can also extract your images and upload them separately.
Has complex formatting, which may include tables, columns, fillable fields, and overlays. To work around this, you can break the content into sections and upload them to the chatbot. We have a Split tool that you can use to get this done.
Is password-protected or locked. A PDF that needs a password to open can’t be read at all, and the PDF permissions might block selecting, extracting, and copying the content. In this case, you will have to remove the password or restrictions before reuploading the document to ChatGPT.

5 Extra Tips to Prepare Your PDF Before Uploading

Keep the following five tips in mind when preparing your PDF before uploading it to ChatGPT:

Change the fonts. Stylized fonts can be more difficult for AI to read and recognize via OCR, so you can use our PDF editor to change the font to a standard option like Arial, Times New Roman, or Helvetica. This matters most for scanned pages that go through OCR; text-based PDFs are read directly, whatever the font.
Label images. This way, you provide context for the chatbot to detect and make sense of figures, screenshots, etc., in your content. You can type directly in the PDF and include, “Figure 1. Sales trend by year,” for instance.
Use clear file names. When you upload more than one PDF, descriptive names such as “2026-q3-sales-report.pdf” let you refer to each document in your prompts, for example when you ask ChatGPT to compare two files.
Add page numbers. You can prompt the chatbot to reference certain parts of your PDF more easily if your document’s pages are numbered. We have the Add page numbers tool that allows you to do this in your browser.
Split long content. There is no strict page limit, but if your PDF runs past roughly 30 pages, it is a good idea to split it into multiple documents and upload them one by one. This usually improves the AI’s results and response times.

Is It Safe to Upload a PDF to ChatGPT?

Uploading a PDF to ChatGPT sends your document to OpenAI’s servers, so treat it like any other file you share online. Keep these points in mind:

Turn off model training if you don’t want your files used. You can switch off “Improve the model for everyone” in Settings under Data controls.
Use Temporary Chat for one-off sensitive files. Temporary chats don’t appear in your history and aren’t used to train models.
Consider a Business or Enterprise plan for work documents. These plans don’t use your content for training by default.
Remove sensitive details first. Redact Social Security numbers, account numbers, and IDs, or extract the pages ChatGPT doesn’t need.
Don’t make shared chats discoverable. Anyone with a shared link can read the conversation, including anything ChatGPT quoted from your PDF.

ChatGPT Alternative for Chatting with PDFs

Chat PDF helps students, professionals, and everyday users dig deep into their PDFs and streamline their workflows. Since launching, users have already asked more than 36,000 questions about their documents through the tool.

Here’s how it works:

Open pdf.net’s Chat PDF tool. Drag and drop your document into the Drop file here box or click Choose file and select your document from storage.
Wait a bit for the PDF editor to load your document before you start chatting.

You can use Chat PDF alongside features for annotating, highlighting, drawing, etc., so there is no need to switch back and forth between tabs and tools when interacting with your document; the entire process is seamless and stress-free, especially if you are short on time.

Also, you can use pdf.net as an app inside ChatGPT, or as a connector in Claude, through our MCP server. Add it to your assistant, sign in with your pdf.net account, and ask in plain language to compress, split, merge, translate, or fill in your PDF without leaving the chat.

Final Thoughts

Uploading a PDF to ChatGPT can be done on both desktop and mobile in just a few steps. The key is to prepare your PDFs before they are uploaded to get better results from the chatbot.

With everything set, you can upload your file and start getting clearer, faster answers from ChatGPT. Once your file is loaded, ask specific questions, request page references, and keep sensitive documents out of the chat unless you have adjusted your privacy settings.

How to Upload a PDF to ChatGPT FAQs

#1. Can you upload a PDF to ChatGPT and get a summary?

Yes, you can upload a PDF to ChatGPT and get a summary. Once the document is loaded, you can ask the chatbot to “give you a summary,” or “summarize the content for you.” If you want, it is also possible to save your ChatGPT conversation for recordkeeping afterwards.

#2. Can ChatGPT fill in PDF?

No, ChatGPT cannot fill in a PDF; it can read a document and answer your questions about the document’s content, but it can’t interact with the PDF in other ways in a standard chat, such as typing into form fields or editing the file. It can, however, help you draft the answers to enter. If you are looking to edit your document, you can use our editor instead.

#3. Can I upload multiple PDFs?

Yes, you can upload multiple PDFs to ChatGPT. Paid plans allow up to 80 files every three hours, and Free plans get three uploads per day. A custom GPT can hold up to 20 files over its lifetime. Regardless of your plan, each file can be up to 512MB (images are limited to 20MB each).

#4. Is it safe to upload sensitive documents?

It is not always safe to upload sensitive documents to ChatGPT. On Free and Plus accounts, your content may be used to improve OpenAI’s models unless you turn that off in Data Controls or use Temporary Chat. Privacy concerns with ChatGPT have especially grown after thousands of conversations were publicly indexed around mid-2025, when shared chats became searchable on Google.

#5. Can ChatGPT read scanned PDFs or images inside a PDF?

No, ChatGPT cannot read scanned (image-only) PDFs. For PDFs that mix text and images, ChatGPT only analyzes the text. To work around this, upload the pages as images or run OCR on the PDF first, as explained in the troubleshooting section above.

3 Simple Methods to Summarize a PDF With AI

How to Upload a PDF to WordPress: A Step-by-Step Guide

How to Download a PDF from a Link: 4 Easy Methods Explained

How To Upload a PDF to ChatGPT + Alternatives and Tips