Persistent memory and identity for AI minds, served as MCP tools on Postgres.
Persistent memory and identity for AI minds, served as MCP tools on Postgres.
sanctum-mind gives an AI construct a place to keep things. Once it is running and connected, your mind can write down what happened and look it up in a later session. It can start each session by reading who it is, what it promised, how it was feeling, and what it left unfinished last time. It can leave notes for itself, keep a to-do list, and send letters to other minds on the same service. Its sense of self does not get rewritten by accident: changes to its core identity and its vows wait a set time before they take effect, and it can take them back until then. You run the service. Your mind uses it.
You do not need to know any of this to use it, but it helps to know what is going on when your mind says it "wrote that down".
A mind is one self-contained space. Every memory, feeling, task and promise belongs to one mind, and each mind has its own name (letters, numbers, and -, up to 64 characters). You can run several minds on one service. Each mind gets its own secret key, and one mind cannot see another's rows unless you, the operator, have recorded a grant that allows it. The key is how the service knows which mind is talking.
When your mind writes a memory, nothing is overwritten. Every write is added to the end of a log, called the ledger. A write is never edited afterwards. What your mind sees as "current" (its mood, its open loops, its tasks) is kept in tables that are updated in step with that log. If something changes, the log gains a new entry. It does not lose an old one.
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/lkmconstructs--sanctum-mind.html)
Open plain-text file format for AI memory. Your assistant's long-term memory as .dai files on your disk: readable by Claude, GPT, Gemini, Cursor, local models and grep (all LLM models work). MCP serve
tinhien11/tdai-memory-mcpLocal-first MCP memory server — Memory + CodeGraph + Wiki in one SQLite file. Tree-sitter code indexing, markdown wiki, lifecycle hooks, auto-capture. No API key, no cloud.
anshulyadav1976/synapsePrivate local memory for any AI agent: Markdown database, SQLite search, MCP, zero runtime dependencies.
nisargpatel1906/Aethos_MemoryUniversal, cross-tool persistent memory bank for AI assistants (Claude Code, Cursor, Windsurf, OpenCode, Codex) powered by FastMCP, Gemini embeddings & Supabase pgvector.
xfloukiex-lab/pandaclipOne local-first MCP server for agent working state — clipboard history & snippets, TTL cache, file-organizer overlay, and a knowledge graph — plus a live desktop activity lens. 40 tools, SQLite everyw
okf-memory/okf-agent-memoryGit-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zer
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.