repository·master·Indexed 25 days ago
https://github.com/pionxzh/chatgpt-exporterA userscript for exporting ChatGPT conversation histories into multiple formats, including Markdown, HTML, Text, JSON, and PNG screenshots. It supports single and bulk exports, importing from official conversations.json files, and administrative actions like archiving or deleting chats. The tool includes a programmatic API for fetching conversations, handling rate limits, and processing raw OpenAI API data into structured formats.
ChatGPT Exporter is a GreasyFork script used to export chat histories from ChatGPT.
You must install a userscript manager like Tampermonkey first. It is available for:
You can install the script via GreasyFork or directly from GitHub:
When you click the "Tout Exporter" (Export All) button, the Exporter les Conversations (Export Conversations) dialog opens.
You can click the download icon to import a conversations.json file, such as the one exported directly from OpenAI.
To export multiple conversations at once, click the "Export All" button to open the "Export Conversations" dialog.
Select your desired format from the dropdown menu in the bottom left corner:
The JSON export format provides the raw content from the OpenAI API endpoint https://chat.openai.com/backend-api/conversation/[id]. It includes conversation metadata such as id, title, create_time, and a mapping object containing the message tree (roles, content, timestamps, etc.).
{
"id": "35a1fa05-e928-4c39-8ffa-ca74f75b509f",
"title": "AI Turing Test.",
"create_time": 1678015311.655875,
"mapping": {
"5c48fa3e-e4ee-4d00-aa66-8fbcb671a358": {
"id": "5c48fa3e-e4ee-4d00-aa66-8fbcb671a358",
"message": {
"id": "5c48fa3e-e4ee-4d00-aa66-8fbcb671a358",
"author": {
"role": "system",
"metadata": {}
},
"create_time": 1678015311.655875,
"content": {
"content_type": "text",
"parts": [
""
]
},
"end_turn": true,
"weight": 1,
"metadata": {},
"recipient": "all"
},
"parent": "9310b90f-d8f0-4be6-bac2-daacddac784f",
"children": [
"4afb9720-3a88-49b1-9309-e2b53d607f34"
]
}
}
}When exporting to JSON, the output is the raw content retrieved from the OpenAI API endpoint https://chat.openai.com/backend-api/conversation/[id].
Key fields in the JSON structure include:
id: The unique identifier for the conversation.title: The title of the chat.create_time: Timestamp of creation.mapping: A dictionary containing conversation nodes, where each node includes message (with author.role, content, etc.), parent, and children IDs.{
"id": "35a1fa05-e928-4c39-8ffa-ca74f75b509f",
"title": "AI Turing Test.",
"create_time": 1678015311.655875,
"mapping": {
"5c48fa3e-e4ee-4d00-aa66-8fbcb671a358": {
"id": "5c48fa3e-e4ee-4d00-aa66-8fbcb671a358",
"message": {
"id": "5c48fa3e-e4ee-4d00-aa66-8fbcb671a358",
"author": {
"role": "system",
"metadata": {}
},
"create_time": 1678015311.655875,
"content": {
"content_type": "text",
"parts": [""]
},
"end_turn": true,
"weight": 1,
"metadata": {},
"recipient": "all"
},
"parent": "9310b90f-d8f0-4be6-bac2-daacddac784f",
"children": ["4afb9720-3a88-49b1-9309-e2b53d607f34"]
}
}
}ChatGPT Exporter supports exporting chat history in the following formats:
https://chat.openai.com/backend-api/conversation/[id])Markdown exports include metadata such as title, source, and author in the YAML front matter.
Use exportAllToJson to export multiple conversations into a single .zip file. Each conversation within the ZIP is saved as an individual .json file. If multiple conversations share the same filename, they are automatically suffixed with a count (e.g., (2).json).
Parameters:
fileNameFormat: The base filename format for individual JSON files.apiConversations: An array of ApiConversationWithId objects._metaList (optional): Metadata for the export.projectName (optional): The name of the project used for the ZIP filename.partIndex (optional): The index of the current part in a batch.totalParts (optional): The total number of parts in the batch.probeApi() to perform a lightweight check to see if the ChatGPT API is currently accepting requests. This is useful for displaying status to users before starting large export tasks. It returns an object containing ok status, retryAfterMs (if rate-limited), and rateLimitHeaders.The getFileNameWithFormat function allows you to generate dynamic filenames using a template string and specific placeholders. The resulting filename is appended with the provided extension.
Available Placeholders:
{title}: The sanitized title (spaces replaced with underscores).{date}: Current date string.{timestamp}: Current timestamp.{chat_id}: The unique ID of the chat (empty if exporting all conversations).{create_time}: ISO string of the creation time (expects Unix timestamp in seconds).{update_time}: ISO string of the update time (expects Unix timestamp in seconds).Options Object:
title: The title to use (defaults to document.title).chatId: The chat ID string.createTime: Unix timestamp in seconds.updateTime: Unix timestamp in seconds.exportToJson to export the current active ChatGPT conversation into a standard JSON format. This format is an array of the raw API response. It requires a fileNameFormat string to determine the filename.exportToTavern to export the current active ChatGPT conversation into the Tavern format. The output is a .tavern file using the application/json-lines (JSONL) MIME type.The exportToPng function captures the current ChatGPT conversation and downloads it as a PNG image. It automatically handles styling (adjusting for dark/light mode), hides UI elements like headers and buttons to ensure a clean screenshot, and attempts to handle large conversations by recursively reducing the scale if the canvas size limit is exceeded.
Requirements:
#thread and [data-testid="conversation-turn-1"]).Returns:
true if the export and download were successful.false if the conversation hasn't started, the element could not be found, or the conversation is too large to export.