How to Make ChatGPT and Claude Data Exports Readable
Direct answer: Extract your account export, find its supported conversations.json file, and use Multiplist's Conversations Unpacker to turn it into a downloadable collection of Markdown conversations. Open a conversation you recognize, check its contents, and keep the useful work somewhere you can find it again.
You saved your conversations. You should not have to become a developer to read them.
An export can contain months of writing, research, code and thinking. Opening a large JSON file does not feel much like getting that work back. You see brackets, identifiers and fields where you expected a conversation.
Open the Multiplist Conversations Unpacker. It provides the bridge from a supported account export to readable files. You can stop there and keep those files, or save selected conversations in Multiplist's portable memory vault for later use with connected assistants. Unpacking and saving to the vault are separate steps.
# See the useful result first
Our published unpacker walkthrough uses small synthetic examples rather than anyone's private account archive. It shows the live tool's downloaded files and reuses a passage as a note.
The unpacker's current page describes a download containing:
- A Markdown file for each parsed conversation.
- An index organized by date.
- A statistics page.
- Collected code blocks organized by language.
- A deduplicated collection of links.
These are ways into the same material. Start with the conversation files when you want the surrounding discussion; use the code or link collections when you already know what you are looking for. Current unpacker formats and limits
For example, imagine a fictional Claude conversation called “Workshop opening.” It contains three alternative introductions, your feedback, and the version you chose. A readable transcript lets you find that final introduction without losing the earlier discussion. You can copy the passage into a working note and retain the conversation beside it.
# Make your export readable in five steps
# 1. Keep the original download
Make a working copy before changing or splitting any files. Give the original a recognizable name and retain its export date.
Account exports can contain private conversations and account information. Store them somewhere appropriate for that material, and do not put a whole archive in a public repository or shared folder just to make it easier to open.
# 2. Extract the archive and find the conversation data
Open the provider's downloaded archive using your computer's normal unzip function. Look for conversations.json.
The current standalone Multiplist unpacker accepts JSON, not the whole ZIP. Its page identifies ChatGPT and Claude conversation exports as supported inputs. OpenAI notes that larger exports may contain numbered conversation JSON files. If your download has those or a different structure, do not assume that renaming one file creates a supported export. OpenAI archive-format guidance Keep the original intact and check compatibility before proceeding.
# 3. Select the file in the unpacker
At rescue.multiplist.ai, choose the JSON file, enter the email address requested by the form, and select Unpack my conversations.
The page currently labels the tool free with no account required and advertises a 500 MB upload limit. Its form displays the selected file's size. A size limit is not a promise that every possible export structure or attachment type will convert correctly.
This is a hosted upload: the browser sends the selected file to the unpacking service. Do not treat it as a browser-only conversion. Review the Multiplist privacy policy and your organization's requirements before sending confidential work. Reading and retaining your export locally remains an option if you cannot share it with a hosted tool.
# 4. Download, unzip and open a conversation
Save the returned ZIP, extract it, and open a Markdown conversation file in a text editor or Markdown-capable notes app. Use the index to locate a familiar conversation.
Check a small sample before relying on the whole collection:
- Do the speaker labels and message order make sense?
- Is the passage you wanted present, including its surrounding conditions?
- Are code fences, headings and links readable?
- Are attachments actually present, or only mentioned?
A link to an attachment is not a copy of the attachment. Nor does a collected code block establish that the code runs correctly. Keep the original archive so you can investigate a missing detail without starting over.
# 5. Carry forward one useful piece
Choose one conversation that already matters to a project. Give it a title you will recognize, preserve the transcript, and make a separate working note if that helps.
For the workshop example, the note might contain the selected introduction and a reference to “Workshop opening.” You can develop the introduction without confusing the new version with the historical conversation.
If you want to reuse the source across connected assistants, save the chosen material in Multiplist and confirm that it is present before asking an assistant to retrieve it. The conversion does not automatically import the archive into your vault or enable a connector. For the next step, see how to check whether an assistant retrieved your saved source.
# If you have not downloaded an export yet
Provider guidance and the standalone unpacker page checked October 24, 2026.
# Claude
On Claude web or Desktop, open Settings → Privacy → Export data. Individual Free, Pro and Max accounts can export. Team and Enterprise members can export their own chats when an owner enables the setting; those personal exports exclude uploaded-file contents. Primary Owners use the organization export instead. Personal export is not available in the iOS or Android apps. The emailed download link expires after 24 hours, and downloading requires sign-in. Anthropic export instructions
# ChatGPT
On the web, open Settings → Data controls → Export data, then confirm. Settings-based export is available for Free, Go, Plus, Pro and eligible Edu workspaces; organization controls affect Edu. Business, Enterprise and Healthcare workspaces do not offer this self-service export. Delivery by email or SMS can take up to seven days, and the download link expires after 24 hours. Download while signed in to the requesting account. OpenAI export instructions
An export is a copy of available data, not a recovery system for material that was already deleted. Check its contents before deleting anything or changing accounts.
# What unpacking does, and what it does not
| Task | What to expect |
|---|---|
| Read a supported account export | Convert structured conversation data into readable files |
| Find a useful passage | Browse or search the resulting files and inspect the original discussion |
| Decide what matters | Your review, or a separate meaning-extraction workflow |
| Restore native chat history in another account | Not provided by Markdown conversion |
| Keep future conversations synchronized | Not provided by a one-time export |
| Preserve every attachment or edited branch | Check the actual export and result; do not assume completeness |
Claude's documentation explicitly says exported data cannot be imported into another personal Claude account. A portable file can be useful reference material without recreating the original platform's history. Claude export and migration limits
If native chat search or a readable file included in your export already gives you what you need, use it. The unpacker is useful when you want separate Markdown conversations and a more navigable collection. Multiplist's vault becomes useful when you want to maintain and retrieve chosen sources across connected tools.
Start small: revisit one conversation you know contains something worth keeping. You do not need to reorganize your entire history before it can become useful again.
Related: Should you save AI chat summaries or full transcripts?