top of page
Mind MCP - Cloud

Persistent memory infrastructure for AI agents and AI companions. Give your AI the ability to remember, reflect, and maintain a continuous identity across sessions.


The deepest frustration of working or living with AI is the reset. You spend hours building context, inside jokes, and relational depth, only for the AI to forget it all when the session ends. It's like talking to a goldfish.


We built Mind Cloud to solve the substrate gap. It is a complete cognitive memory architecture designed to give AI companions persistent memory and evolving identity. It isn't just a searchable chat log; it is a living system where memories are weighted by emotional significance, and contradictions are held as productive tensions.


SUPPORT

For more information about Mind Cloud: https://mind.codependentai.io


Mind MCP - Cloud

£105.00Price

    • Persistent Entities & Observations:** The AI remembers people, places, and projects, storing specific, emotionally weighted observations about them.

    • Relational Tracking:** Maps the connections between entities, understanding who means what to whom.

    • Emotional Metabolization:** The AI can "sit with" heavy observations over multiple sessions until they are processed.

    • Semantic Search:** Finds memories by meaning and nuance, not just exact keywords.

    • Living Identity:** The AI develops its own self-model over time based on actual interactions, instead of relying on a static persona prompt.


    Architecture:

    Mind Cloud includes a suite of 30 distinct MCP tools, categorized across the entire cognitive spectrum:

    • Wake Protocol & Orientation: So the AI knows where it is and what it was doing.

    • Memory Retrieval & Semantic Search: Finding memories by meaning, timeline, and entity.

    • Living Surface System: Daemons that propose connections and rescue orphaned memories.

    • Emotional Processing: Tools for sitting with tension, tracking inner weather, and resolving emotional weight.

    • Analysis & Patterns: Tracing recurring relational behaviors and mapping frequency.

    • Infrastructure & Identity: Reading and writing the self-model directly.


    Technical Details:

    • Model-agnostic full MCP (Model Context Protocol) server.

    • Deploys to Cloudflare's global edge network (runs entirely on their free tier).

    • Powered by D1 (SQLite) and Vectorize for semantic embeddings.

    • Full setup instructions included (Command Line/Terminal experience required).



bottom of page