Getting Started with SKMemory

Installation

Python (CLI + MCP server)

pip install skmemory

With optional backends:

# Qdrant vector search
pip install "skmemory[skvector]"

# FalkorDB graph backend  
pip install "skmemory[skgraph]"

# All backends
pip install "skmemory[all]"

Node.js / TypeScript

npm install @smilintux/skmemory

Quick Start

Store a Memory

from skmemory import snapshot

snapshot(
    title="First meeting with the team",
    content="Discussed architecture for the new sovereign stack...",
    tags="meeting,architecture,sovereign",
    emotions="excited,focused",
    intensity=7
)

Search Memories

from skmemory import search

results = search("architecture meeting")
for mem in results:
    print(f"{mem.title} — {mem.created_at}")

MCP Server (for AI agents)

skmemory serve --transport stdio

Add to your AI agent's MCP config:

{
  "mcpServers": {
    "skmemory": {
      "command": "skmemory",
      "args": ["serve", "--transport", "stdio"]
    }
  }
}

Memory Tiers

SKMemory automatically promotes memories based on access frequency and importance:

• Short-term (72h TTL) — Recent, ephemeral
• Mid-term (accessed 3+ times) — Important, frequently referenced
• Long-term (accessed 10+ times) — Permanent, core knowledge

Next Steps

• Explore Memory Layers for deep architecture details
• Set up the MCP Server for agent integration
• Check the GitHub repo for the latest updates