Export attachments from ChatGPT and DeepSeek, not just the text.
Copy a conversation out by hand and you keep the commentary and lose the material — the spreadsheet you uploaded, the screenshot you pasted, the chart the model drew back. Extry collects every file in the conversation and writes them next to the transcript.
- conversation.md18.4 KBmarkdown
- files/budget-q3.xlsx88.6 KBattachment
- files/diagram-v3.png480.2 KBattachment
- files/notes-scan.pdf212.4 KBattachment
What comes out
Two things people usually lose
Extry walks the whole conversation, top to bottom, and picks up both of them.
Files you uploaded
Spreadsheets, PDFs, CSVs, documents — saved under their original names.
Files ChatGPT produced
Generated images, charts, and downloads it offered you mid-conversation.
Everything lands in a files/ folder next to the transcript, and the links inside the transcript point at it. Open the archive in a year and it still resolves.
Why it matters
Why the transcript alone is not enough
A working conversation is mostly not prose. It is the file you fed the model and the artifact it handed back. A transcript without them reads like one half of a phone call: the reasoning survives, the subject of the reasoning does not.
This matters most in exactly the cases where you bothered to export: a review you need to reconstruct, an analysis you have to hand to someone else, a decision you will be asked to justify.
How it works
Three steps, one archive
- Open the conversation in ChatGPT or DeepSeek and click the Extry button on the page.
- Pick the formats you want and leave attachments switched on.
- Press Download this chat and get the transcript and every file in one archive.
No account, no upload step. The extension reads the page you already have open.
Processing
Where each part runs
| What | Where it happens |
|---|---|
| Export to every format | In your browser |
| Attachments and images | In your browser |
| Conversation content | Not sent anywhere |
| Usage statistics | Not collected |
| Typeset PDF (later, opt-in) | On a server, file deleted right after |
Questions
About attachments
Are generated images included?
Yes. Images ChatGPT produced are saved alongside the files you uploaded.
What happens to file names?
Original names are kept. Duplicates get a numeric suffix so nothing overwrites anything.
Do the links in the transcript still work?
Yes. Links point at the files/ folder inside the archive, so the export stays self-contained.
Is there a size limit?
Past roughly 1.5 million characters of transcript the HTML and Word copies are skipped, because rendering them in the page stops being safe; Markdown, text, JSON, and the attachments still come out.
More
Other ways out
The ZIP is one of nine ways out. Markdown, Word, printing to PDF, the images, the bulk export, the main page.