How to Export Your ChatGPT Conversation History (Complete Guide)

ChatGPT is one of the few AI platforms that offers a genuine data export. This gives you a complete, portable copy of everything you've generated — all conversations, including messages, timestamps, and model information.

The export is underused. Many users don't know it exists, and those who do often don't know what to do with the resulting files. This guide covers the full process: requesting the export, understanding what you receive, and making that data useful.

Requesting the ChatGPT data export

The export process takes about five steps:

1. Sign in to ChatGPT at chatgpt.com
2. Click your profile icon in the bottom-left corner
3. Select Settings
4. Navigate to Data controls (in the left panel)
5. Click Export data — a confirmation dialog will appear
6. Confirm the export request

OpenAI will send an email to the address associated with your account. The email contains a download link. The link typically arrives within a few minutes for small accounts, and may take a few hours for accounts with very large conversation histories.

Important: The download link expires — usually within 24 hours. Download the file as soon as you receive the email. If the link expires before you download it, you'll need to request a new export.

What's included in the export

The download is a zip archive containing several files:

conversations.json — The complete machine-readable export. This is the most valuable file. It contains:

• Every conversation, including its unique ID and title
• Every message in every conversation, with timestamps
• Which model was used for each response (GPT-3.5, GPT-4, GPT-4o, etc.)
• The full text of both your messages and ChatGPT's responses
• Conversation creation and update timestamps

chat.html — A human-readable HTML version of your conversations. Open this in any browser. It renders your conversations in a chat-like format that's easy to read. This is what most users will want for browsing their history.

message_feedback.json — Any thumbs up/down ratings you gave to responses.

model_comparisons.json — Data from any A/B model comparison experiments you participated in.

user.json — Your account information (name, email, account creation date).

How to search the exported data

Option 1: Browser search in chat.html

1. Open the exported chat.html file in any browser (double-click it or drag it to a browser window)
2. Use Ctrl+F (Windows) or Cmd+F (Mac) to search across all conversations
3. The search finds matches in conversation titles and message content

This is the easiest option and requires no technical knowledge. The limitation is that very large exports may slow down the browser — if you have thousands of conversations, the HTML file can be several hundred megabytes.

Option 2: Text search in conversations.json

For users comfortable with text editors:

1. Open conversations.json in a capable text editor (VS Code, Sublime Text, or similar)
2. Use the editor's find function (Ctrl+F) to search for keywords

JSON is raw text, so keyword search works directly. The context around matches shows the full message. The downside is that JSON is not formatted for reading — the text is dense and requires some mental parsing.

Option 3: Import into a search tool

Several approaches allow you to make the export more permanently searchable:

• Obsidian or Notion: Convert the JSON to markdown format (scripts for this exist on GitHub) and import into your note-taking tool
• SQLite: Parse the JSON into a local database for faster queries
• LLMnesia: For ongoing indexing rather than point-in-time exports, LLMnesia indexes your active ChatGPT sessions automatically, so you don't need to manage exports at all

Limitations of the ChatGPT export

No images: If ChatGPT generated images in a conversation (via DALL·E), the export includes a reference to the image but not the image file itself.

No browsing results: For conversations where ChatGPT used web browsing, the export includes the text response but not the source pages or links that were retrieved during the browse.

Memory not included: ChatGPT's "memory" feature (where ChatGPT remembers facts about you between conversations) is stored separately. The conversation export does not include memory contents — request a separate export for that.

Shared conversations: If someone shared a conversation link with you and you accessed it, that conversation does not appear in your export — only conversations from your own account are included.

Why exporting is different from ongoing indexing

The export is a snapshot: it captures your conversation history at the moment of export. Conversations you have after the export date are not included until you run another export.

For ongoing access to recent conversations, a browser extension like LLMnesia provides continuous indexing — every conversation is captured as you have it, without needing to request and manage periodic exports.

The two approaches are complementary:

• Export: Useful for creating a permanent backup, recovering old conversations that predate browser extension installation, or migrating data
• Ongoing indexing: Better for regular access to recent conversations, cross-platform search, and not having to manually manage export files

How to use the export for conversations before you installed LLMnesia

If you've recently installed a conversation indexing extension and want to make your pre-installation conversation history searchable:

1. Export your ChatGPT data using the steps above
2. Open the exported chat.html file and browse through your older conversations
3. For any valuable conversations, opening them on the ChatGPT website will trigger indexing (LLMnesia indexes conversations as you view them)

This doesn't automatically index your entire history from the export file, but it lets you manually identify and index the conversations that matter most to you.

Protecting your exported data

Your conversation export contains everything you've said to ChatGPT. Handle it accordingly:

• Store the zip file in a location with appropriate access controls (not a public cloud folder)
• If you share a device, keep the file in a location others can't access
• Consider encrypting the archive if it contains sensitive professional or personal content

The HTML format is convenient for reading but is entirely plain text — easy to open and read, but also easy for anyone else to read.