mempalace

name	mempalace
description	MemPalace is a local-first AI memory system. It stores conversation context, Use when this capability is needed.
metadata	{"author":"bunkerlab-net"}

MemPalace

MemPalace is a local-first AI memory system. It stores conversation context, project knowledge, and code insights in an embedded SQLite database accessible via MCP tools or the mempalace CLI. No cloud services or API keys required.

Architecture

Wings (top-level categories, e.g. projects, people, domains)
  +-- Rooms (sub-topics within a wing)
        +-- Drawers (verbatim memory chunks)

Tunnels connect related rooms across different wings.

When to Use MemPalace

• Remember context across sessions: call mempalace_search when the user references something that may have been discussed before.
• Wake up at the start of a session: check mempalace_status and the L0/L1 context from mempalace wakeup to orient yourself.
• File important facts: use mempalace_add_drawer to store decisions, preferences, milestones, and key facts the user wants to be remembered.
• Mine projects: run mempalace mine <dir> to ingest a project directory or conversation exports into the palace.

MCP Tools Quick Reference

Search and recall

mempalace_search(query, wing?, room?)     -- keyword search with BM25 ranking
mempalace_list_wings()                    -- list all top-level wings
mempalace_list_rooms(wing)                -- list rooms within a wing
mempalace_get_taxonomy()                  -- full wing/room/drawer tree
mempalace_traverse(room, wing?)           -- graph traversal from a room
mempalace_find_tunnels(wing1, wing2)      -- cross-wing connections
mempalace_follow_tunnels(wing, room?)     -- follow tunnels from a location

Write

mempalace_add_drawer(wing, room, content, source_file?)   -- store a memory
mempalace_update_drawer(id, content)                      -- update a memory
mempalace_delete_drawer(id)                               -- remove a memory
mempalace_get_drawer(id)                                  -- fetch one drawer
mempalace_list_drawers(wing?, room?, limit?, offset?)     -- paginated listing

Knowledge Graph

mempalace_kg_query(entity?, predicate?, object?)    -- query KG triples
mempalace_kg_add(subject, predicate, object, ...)   -- assert a KG fact
mempalace_kg_invalidate(id)                         -- retract a KG fact
mempalace_kg_timeline(entity)                       -- entity history
mempalace_kg_stats()                                -- graph statistics

Diary

mempalace_diary_write(content, date?)    -- write a diary entry
mempalace_diary_read(date?, wing?)       -- read diary entries

Palace topology

mempalace_graph_stats()    -- room/tunnel/wing-edge counts and connectivity
                              (distinct from `mempalace_kg_stats` above,
                               which counts knowledge-graph triples)

Slash Commands

Command	What it does
/mempalace:init	First-time setup wizard
/mempalace:mine	Guide through mining a directory
/mempalace:search	Guide through searching memories
/mempalace:status	Show palace overview
/mempalace:help	Show all commands and MCP tools

CLI Commands (fallback when MCP is unavailable)

mempalace init <dir>                 Scan project, detect rooms, write config
mempalace mine <dir>                 Mine project files
mempalace mine <dir> --mode convos   Mine conversation exports
mempalace search "query"             Keyword search
mempalace wakeup                     Print L0+L1 wake-up context
mempalace status                     Palace stats overview
mempalace compress                   AAAK dialect compression
mempalace split <dir>                Split large transcript files
mempalace repair                     Rebuild inverted index
mempalace mcp                        Run MCP server (stdio)

Auto-Save Hooks

Two hooks auto-file memories during Claude Code sessions:

• Stop (mempalace hook --hook stop --harness claude-code): runs after every N exchanges, filing a diary checkpoint and optionally mining the current transcript directory.
• PreCompact (mempalace hook --hook precompact --harness claude-code): runs before context compaction, mining the transcript so memories land before the window closes.

Hooks are pre-configured by the plugin's hooks.json. Manual install instructions are in hooks/mempal_save_hook.sh and hooks/mempal_precompact_hook.sh.

---

Source: bunkerlab-net/mempalace — distributed by TomeVault.