# @memstack/mcp

Source: https://glimind.com/server/npm%2F-memstack%2Fmcp
Registry id: `npm/-memstack/mcp`
Status: unknown

## Install

```json
{
  "mcpServers": {
    "memstack-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@memstack/mcp"
      ]
    }
  }
}
```

Claude Code:

```sh
claude mcp add memstack-mcp -- npx -y @memstack/mcp
```

## Tools (18)

- `memory_compile_context`: Assemble token-budgeted LLM-ready context
- `memory_delete`: Delete a single memory
- `memory_delete_many`: Delete multiple memories by ID
- `memory_dry_run_prune`: Preview what would be pruned
- `memory_export`: Export a memory snapshot for backup or migration
- `memory_get`: Get a single memory by ID
- `memory_health`: Check storage, LLM, and embedding connectivity
- `memory_import`: Import memories from a snapshot produced by memory_export
- `memory_merge`: Merge multiple memories into one
- `memory_process`: Store a memory with auto-enrichment (importance, tags)
- `memory_prune`: Remove stale or low-importance memories
- `memory_purge_actor`: Delete all memories for an actor
- `memory_retrieve`: Retrieve memories by query, strategy, time range
- `memory_stats`: Memory diagnostics (counts, types, importance)
- `memory_store`: Store a memory
- `memory_store_batch`: Store multiple memories in one call with batched embeddings
- `memory_summarize`: Compress old interactions via LLM
- `memory_touch`: Bump a memory's recency without changing content

Measured by Glimind (https://glimind.com): MCP handshake + tool list only, no tool is ever called. CC BY 4.0.
