Save a ChatGPT conversation as a PDF

Why a normal "print the page" fails here

ChatGPT renders a conversation as a virtualized list: older messages are removed from the DOM as you scroll. Printing or screenshotting the page therefore captures only the fragment currently loaded—usually with the site's interface around it and no reliable order.

How the complete export works

When you start an export from an open conversation, Web Paperize builds the document from the conversation itself rather than the on-screen DOM:

1. Using your existing logged-in session, the extension requests the conversation you are viewing, directly from ChatGPT/OpenAI—only for this export.
2. Messages are assembled in order along the active branch; internal notes, tool calls, and hidden entries are left out.
3. Citations become numbered references with a source list per message; conversation images are embedded at their original position; attachments are listed with name, type, and size.
4. Chrome renders the result as a searchable, paginated PDF with working links.

If the conversation cannot be fetched—for example, you are not logged in—Web Paperize says so instead of quietly exporting an incomplete page. A context-menu entry lets you export the currently loaded content as a normal webpage whenever you prefer that.

What this means for your data

Complete-conversation export is the only Web Paperize feature that requests conversation data directly from ChatGPT/OpenAI, and it runs only when you ask for it:

• Requests go directly from your browser to ChatGPT/OpenAI, using the session you already have, solely to build the PDF you requested.
• Web Paperize has no backend server. Your access token stays in memory for the request and is never logged, stored, or sent to the developer.
• Ordinary webpage exports are processed locally in Chrome—Web Paperize does not send page content to a Web Paperize server or any unrelated third party. The page's own normal requests while it loads or scrolls stay between you and that site.

Details are in the privacy policy. Web Paperize is an independent extension and is not affiliated with OpenAI.

Good fits

• Archiving a research or troubleshooting thread you want to keep searching later.
• Sharing a conversation as one document, with sources visible and clickable.
• Keeping generated code blocks, tables, and file attachments together in context.

Known limits

• The export uses the conversation's active branch; edited or abandoned branches are not merged in.
• Very long conversations produce very long PDFs—expect large files and, on some viewers, sparse pages around big code blocks.
• Sandbox files referenced by a conversation are listed, not embedded.