Features

Add analytics that summarize the user's memory patterns and provide insight into what their stored context says about their interests, focus areas, and changes over time. Possible outputs: Recent focus areas Recurring themes Topics that have grown or faded over time A sanitized public-facing summary with no personal information Optional year-end recap-style report Important caution: This feature should avoid drifting into invasive or overly psychological analysis. It should focus on useful, user-controlled summaries rather than making sensitive personal judgments. Users should be able to opt in or out. Why it matters: Memory analytics could help users understand their own work patterns and areas of focus. It could also create a shareable, marketing-friendly product moment (similar to Spotify Wrapped) that makes the product feel more alive and personal.

Allow Membase to store images and retrieve them through MCP. Use cases: Product images and logos Whiteboard photos Diagrams Website screenshots Visual references agents can analyze later Desired behavior: AI agents should be able to retrieve images visually and analyze them for context when needed. Images should be persistently stored in the user's Membase account and retrievable in future conversations. Future possibility: Long term, video storage and retrieval could also be valuable, though likely more expensive to support. Why it matters: A lot of useful context is visual. Supporting image retrieval would make Membase useful for design, product work, education, research, and visual documentation workflows. It would also address a gap that text-only memory systems cannot fill.

Memory and wiki currently feel like separate systems. Memories update automatically, while wikis feel more structured but more manual. User problem: Users may second-guess where information should live: memory or wiki. This creates friction and makes the system feel less unified. Desired behavior: Membase could provide a more unified system where memories and wiki content live in the same overall view but can be filtered, tagged, or organized by type. Why it matters: A unified view would make Membase easier to understand and reduce the mental overhead of deciding whether something belongs in memory or wiki. The two systems serve complementary purposes, but the seam between them should feel invisible to the user.

Current behavior: All stored memories appear to be retrieved through the same MCP tools, regardless of how sensitive they are. User problem: As users store increasingly personal context in AI memory systems, they need stronger control over sensitive information. Treating all memories the same creates trust and safety concerns. Desired behavior: Create a separate MCP tool (e.g., get_sensitive_memory) that sits behind a permission layer, giving users explicit control over when an AI can access their most private information. Desired permission model (similar to Claude MCP tool permissions): Always ask for permission Always allow Never allow Sensitive data classification - privacy-preserving approaches: Option 1: User-controlled tagging Users manually mark a memory as sensitive when storing or editing it. Simple to understand, gives full control, requires no server-side content analysis. Option 2: AI-side lazy marking When an AI retrieves a memory during normal use and identifies it as sensitive, it calls a tool like mark_as_sensitive. Membase only stores metadata ("this memory key = sensitive") without reading the content or knowing why it was flagged. Classification happens entirely in the AI layer. Ideal solution: Use both approaches together - users can proactively tag sensitive memories, and the AI can flag sensitive memories when it encounters them during normal use. Why it matters: This would be a strong trust and safety differentiator for Membase, especially as users store more private and personal context in AI memory systems.