chatgpt-import

What This Skill Does

The chatgpt-import skill is a powerful migration and archival tool designed to bridge the gap between your legacy ChatGPT conversation history and the OpenClaw ecosystem. By transforming raw exported JSON data from OpenAI into a structured, searchable SQLite database, this skill allows OpenClaw to reference your past insights, technical snippets, and creative brainstorming sessions directly within its memory search module. Rather than relying on sporadic manual copy-pasting, this skill automates the ingestion process, ensuring that your long-term memory is fully indexed and contextually aware of your prior interactions.

Installation

To install this skill, run the following command in your terminal: clawhub install openclaw/skills/skills/samdickson22/chatgpt-import

Ensure you have the necessary Python dependencies installed as specified in the source repository. Once the skill is installed, follow the documented workflow: export your data from ChatGPT, run the convert_chatgpt.py script to generate markdown files, and finally use bulk_embed.py to create the SQLite memory file. Update your OpenClaw config.yaml to include the path to your new .sqlite database and restart the gateway to initialize the index.

Use Cases

• Project Continuity: Recall specific technical decisions or codebase explanations discussed in ChatGPT months ago when starting a new iteration of a project.
• Knowledge Management: Create a private, local knowledge base of curated answers, coding solutions, and brainstorming notes that you can query anytime.
• Migration: Transition your primary AI workspace from a cloud-only environment to an integrated, local OpenClaw experience without losing historical context.
• Research Archival: Maintain a searchable database of long-form research conversations for quick reference during current technical documentation tasks.

Example Prompts

1. "Search my chat history for the explanation of the custom middleware we designed for the legacy authentication project."
2. "Find the brainstorming notes regarding the project architecture from my conversations last spring."
3. "Look through my imported ChatGPT logs and summarize the common bugs we encountered while deploying the Python backend."

Tips & Limitations

This process requires an active OpenAI API key for generating vector embeddings. While the cost is negligible (typically less than $0.20 for thousands of conversations), be mindful of sensitive data within your exported files; if you have stored passwords or tokens in your chat history, scrub them manually before running the embedding script. The process is one-way, meaning updates to your OpenAI account will not automatically sync unless you re-run the import process. Ensure your SQLite file is stored locally to maintain privacy and offline accessibility.