ChatGPT-OpenWebUI Import Tool

A simple Python tool to import data exported from ChatGPT (currently only "memories") and sync them with OpenWebUI.

---

Features

• Extract ChatGPT memories from the official export .json.
• Compare & sync: detects which memories to add or delete on OpenWebUI.
• Works with multi-threading for faster processing.

---

Prerequisites

• Python 3.9+ (might work with older versions, but 3.9+ is recommended).
• pip to install dependencies.

Dependencies

pip install requests tqdm python-dotenv pydantic

---

Setup

1. Export your ChatGPT data:

   • Go to ChatGPT → Settings → Data controls → Export all data.
   • You will receive an email from OpenAI with a download link to a .zip.
   • Extract the .zip – locate the conversations.json file.

2. Create a .env file in the project folder (or set them as environment variables). The file should contain:

   OPEN_WEBUI_URL=http://your-open-webui-url
   USER_ID=your_user_id
   JWT_TOKEN=your_jwt_token
   

   • OPEN_WEBUI_URL: Base URL to your OpenWebUI instance (e.g. http://localhost:8080).
   • USER_ID: Your user ID in OpenWebUI.
   • JWT_TOKEN: Your OpenWebUI JWT auth token.

3. Place your conversations.json in the same directory as the script (or modify the script paths accordingly).

---

Usage

1. Install dependencies (if not done already):

   pip install -r requirements.txt

   Or individually:

   pip install requests tqdm python-dotenv pydantic

2. Run the script:

   python import_tool.py

   • By default, it reads from conversations.json and outputs to extracted_memories.jsonl.

3. What it does:

   • Extract memory messages from conversations.json.
   • Compare them to the existing memories in OpenWebUI.
   • Upload new or out-of-date memories.
   • Delete older duplicates on the server if there's a newer local memory.

You’ll see console output showing progress bars and any errors.

---

Notes

• The script currently only deals with "memories" (recipient == "bio") in ChatGPT exports. Other data might be added in the future.
• Check the extracted_memories.jsonl file after running to see exactly which memories were parsed from the conversation data.

---

Contributing

Feel free to open issues or PRs if you’d like to add features (e.g., handling different parts of the ChatGPT export besides memories).

---

License

This project is licensed under the MIT License.