How to Export Your ChatGPT History and Keep a Usable Copy
Direct answer: Request a data export from ChatGPT's settings, download the archive and keep the original. If you want separate readable conversations, use Multiplist's Conversations Unpacker with a supported conversation JSON file, then inspect the result before relying on it.
You may have a draft, research trail or useful piece of code buried in a long conversation. Getting your history back should give you a way into that work, rather than another file you never open.
Open Multiplist's Conversations Unpacker once your download is ready. It turns supported ChatGPT conversation exports into Markdown files. Keep them locally, or choose material to save in Multiplist's portable memory vault for later retrieval by connected assistants. Conversion and vault storage are separate steps.
OpenAI documentation and the unpacker page checked October 26, 2026.
# Request and download your export
On ChatGPT web:
- Sign in and open your profile menu.
- Choose Settings → Data controls.
- Select Export under Export data, then confirm.
Delivery can take up to seven days. Watch the account's email or SMS messages; the download link lasts 24 hours. Download while signed in to the requesting account. OpenAI also provides a Privacy Portal route.
Wait for a pending request rather than repeatedly submitting it. If delivery exceeds seven days, contact OpenAI Support. Current export instructions
An expired link needs a new request. If an unexpired link fails on a phone, try a computer using the same account. The download is a ZIP containing conversation and account data.
# If the export option is missing
Settings export supports Free, Go, Plus, Pro and eligible Edu workspaces, but not Business, Enterprise or Healthcare workspaces. For managed work, ask the workspace owner about the appropriate data-access route. Export availability
Edu has a specific boundary: an administrator must enable the member's export permission, and the workspace must have no data residency configuration. Members export their own data, not everyone else's. A missing button may reflect those controls rather than a broken account. Edu export requirements
Do not use a converter or browser extension to work around your organization's restrictions.
# Open the archive before choosing a tool
Keep the downloaded ZIP unchanged and make a working copy. Extract that copy into a clearly named folder.
OpenAI documents conversations.json, or numbered conversation JSON files in larger exports. These are structured data files, not necessarily comfortable reading material. Conversation-file formats
Look at what your archive actually contains:
- Can you identify a conversation you wanted?
- Are important files present as files, or only referenced by name or link?
- Is the conversation data one file or several?
- Can you already read what you need with the tools you have?
Keep this folder private. An account export can include much more than the single conversation you intended to reuse.
If a readable view in your archive or ChatGPT's own history already meets your need, use it. You do not have to convert everything.
# Turn a supported conversation file into readable Markdown
The current Multiplist unpacker accepts supported ChatGPT conversation JSON. Extract the provider ZIP first; do not submit the whole archive in place of the JSON.
Choose the file, enter the email requested by the form and select Unpack my conversations. Download and extract the result, then open a familiar Markdown conversation. The existing unpacker demonstration shows this workflow using synthetic examples.
If your archive contains numbered files or a different layout, check compatibility first. Renaming a file does not change its structure, and this guide does not establish that every multipart export is supported.
This is a hosted upload. Review Multiplist's privacy policy and the requirements for your material before sending it. Keeping the archive locally is an option.
For the full conversion walkthrough, see How to Make ChatGPT and Claude Data Exports Readable.
# Keep one useful result with its conversation
Imagine a fictional ChatGPT conversation titled Autumn workshop invitation. You tried several versions, changed the audience and finally chose a paragraph.
After conversion:
- Find that conversation in the downloaded collection.
- Compare the speaker labels, order and selected paragraph with the original.
- Copy the chosen paragraph into your working document.
- Keep the transcript beside it, so the earlier context remains available.
- Preserve any needed attachment separately after checking the actual file.
That is already a useful result. You have a working paragraph and the discussion that explains it, without having to reorganize your whole archive.
If you later want another assistant to continue the work, save the selected source in Multiplist, connect and authorize the assistant, and ask it to retrieve that source. Check that it found the intended conversation before building on it. Unpacking alone does not import your archive into the vault or synchronize future chats.
# A reference copy does not rebuild an account
OpenAI describes uploading exported conversation JSON into a new conversation in another personal account as reference material. It does not recreate separate historical chats, restore the old sidebar, merge accounts or transfer subscriptions and settings. Larger files may need separate uploads or a smaller selection. OpenAI's transfer limits
Use that native reference route if it fits your task. Multiplist's role is keeping chosen material available as a shared source for connected tools, not rebuilding the original ChatGPT account.
Before changing accounts or discarding anything, open the files you care about. A successful download is a starting point; a checked, readable conversation is something you can actually use.
Related: Export your Claude history · Check whether your assistant retrieved the saved source