Skip to main content
AI Chat Importer

Why Does Your ChatGPT Export Have Multiple Files? (conversations-000.json Explained)

If your ChatGPT export contains conversations-000.json, conversations-001.json, and a manifest file instead of one file, here's what changed and how to import it correctly.

RM
By R. Miller · AI Chat Importer

If you've just requested your ChatGPT data export expecting a single file, you might be looking at something unexpected: a handful of oddly numbered files instead — conversations-000.json, conversations-001.json, maybe more — plus an extra file that doesn't match either. Nothing has gone wrong. This is a real, known change to how ChatGPT structures exports, and it's easy to import correctly once you know what you're looking at.

What Changed

ChatGPT exports used to produce a single conversations.json file containing every conversation in your account as one large JSON array. As accounts and conversation histories have grown, larger exports are now sometimes split into multiple numbered files instead — each one a chunk of your full conversation history — along with a manifest file that indexes how the pieces fit together.

Not every export is affected. If you have a smaller conversation history, you may still receive the traditional single conversations.json file. The split behavior appears tied to export size, not something you can control directly.

What Each File Actually Contains

  • conversations-000.json, conversations-001.json, etc. — Each is a JSON array of conversation objects, structurally identical to what used to be bundled into one file. Nothing about the individual conversation data has changed — it's simply been divided across multiple files.
  • The manifest file — An index file that lists which numbered files exist and how they relate to the full export. You don't need to open or edit this yourself.
  • chat.html — Still included alongside, as before, offering a basic browsable view of your conversations outside of any importer tool.

How to Import a Multi-File Export

This is the part that trips people up: it's natural to select all the numbered files at once, or drag the whole folder into an import window, expecting it to work like a single file would. Right now, that's exactly what causes an import error.

The correct approach is to import each numbered file one at a time:

  1. Unzip your ChatGPT export if it arrived as a ZIP file.
  2. Open AI Chat Importer (Desktop App or the free web app).
  3. Click the import button and select conversations-000.json. Let it finish importing.
  4. Repeat for conversations-001.json, and any further numbered files, importing each one individually.
  5. Smart Import automatically merges everything into a single searchable archive and flags any duplicate conversations along the way — you don't need to worry about ending up with fragmented or repeated data across the files.

It takes a few extra clicks compared to a single-file import, but the result is the same complete, unified archive either way.

We're working on adding proper multi-file selection support so this becomes a single step in a future update.

Why This Doesn't Affect Claude, Grok, or DeepSeek Exports

This split-file behavior is specific to ChatGPT's export format. Claude and DeepSeek exports currently use simpler, single-file structures, and this guide doesn't apply to them — see our export guides for Claude and DeepSeek if you're working with those instead.

Keep Everything Searchable in One Place

However your ChatGPT export arrives — one file or several — the result once imported is the same: a single, searchable, local archive of your full conversation history. AI Chat Importer processes everything on your own device, with nothing uploaded anywhere.

Get the Desktop App — £29 one-time for permanent, offline, searchable access to your full ChatGPT history.

Prefer to try it first? The free web app handles the same import process, no account required.

FAQ

Why does my ChatGPT export have multiple files instead of one? Larger ChatGPT exports are now sometimes split into multiple numbered files plus a manifest, rather than a single conversations.json. This is tied to export size and isn't something you control directly.

Can I import all the numbered files at once? Not currently — each file needs to be selected and imported individually. Selecting or dragging in multiple files at once causes an import error. Smart Import merges them into one archive automatically once each is imported in turn.

What is the manifest file for? It's an index file listing how the numbered conversation files relate to each other. You don't need to open or use it directly — just import the numbered conversations-*.json files one at a time.

Does this happen with Claude, Grok, or DeepSeek exports too? No — this is specific to how ChatGPT structures larger exports. Claude and DeepSeek currently produce single-file exports.

Will multi-file selection be supported in the future? We're working on adding support for selecting multiple files at once, so this becomes a single step rather than several. For now, importing one at a time is the reliable method.

Related Guides