Quoting claude.com/import-memory
Summary
Anthropic's "import your memories to Claude" feature, accessible via claude.com/import-memory, is presented as a specific prompt designed to extract a user's stored data. Dated March 1, 2026, this prompt instructs the AI to list all memories and learned context about the user in a single code block. The required output format is "[date saved, if available] - memory content," with an emphasis on verbatim preservation of user input. The prompt explicitly requests details such as response instructions (tone, format, style), personal information (name, location, job, family, interests), projects, goals, recurring topics, tools, languages, frameworks used, preferences, corrections, and any other stored context. It strictly forbids summarization, grouping, or omission of entries, and requires a confirmation of completeness after the code block.
Key takeaway
For AI product managers designing data export features, your implementation should prioritize explicit user control and comprehensive data retrieval. Provide a clear, structured prompt that allows users to extract all stored personal data and learned context, ensuring verbatim preservation and a machine-readable format. This approach fosters transparency and empowers users with full data portability, which is crucial for trust and compliance.
Key insights
Users can prompt Claude to export all stored personal data and conversational context.
Principles
- Data portability is a user right.
- Transparency builds user trust.
Method
A specific prompt instructs the AI to list all stored memories and context in a structured, verbatim code block, followed by a completeness confirmation.
In practice
- Use explicit prompts for data export.
- Specify output format for machine readability.
Topics
- Claude AI
- Data Export
- User Memory
- AI Prompts
- Personal Data
Best for: CTO, VP of Engineering/Data, Director of AI/ML, Prompt Engineer, AI Chatbot Developer, AI Ethicist
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by Simon Willison's Weblog.