Self-hosted, local-first MCP memory server for AI agents. Sub-20ms recall at 10k memories, hybrid semantic + exact-syntax search (RAG), code & concept knowledge graphs (158 languages). All local in SQLite: no cloud, no vector DB, no API keys. Works with Claude Code, Codex, Cursor, Gemini & any MCP client.
View on GitHub →
🔍 Audit Wallet Slippage →
Quick Install
Copy the config for your editor. Some servers may need additional setup — check the README.
Claude Desktop
Claude Code
Cursor
Add to claude_desktop_config.json:
{
"mcpServers": {
"marm-memory": {
"command": "uvx",
"args": [
"marm-memory"
]
}
}
}📋 Copy
Run in terminal:
claude mcp add marm-memory uvx marm-memory📋 Copy
Add to .cursor/mcp.json:
{
"mcpServers": {
"marm-memory": {
"command": "uvx",
"args": [
"marm-memory"
]
}
}
}📋 Copy
Or install with pip: pip install marm-memory
README Excerpt
<div align="center"> <picture> <img src="https://raw.githubusercontent.com/Lyellr88/marm-memory/MARM-main/assets/marm-logo.png" alt="marm-memory - persistent local memory server for AI agents (Model Context Protocol)" width="800" height="200"> </picture> <h1 align="center">MARM: Local-First Persistent Multi-Agent Memory Layer for MCP Clients v2.20.0</h1>
Tools (20)
COMPACTION_ENABLED COMPACTION_STAGING_TTL_HOURS COMPACTION_TRIGGER_COUNT CONCEPT_BUILD_ROW_CAP CONSOLIDATION_ENABLED CONSOLIDATION_THRESHOLD FTS_CANDIDATE_LIMIT GRAPH_ENABLED HYBRID_SEARCH_TEXT_WEIGHT MARM_API_KEY MARM_CONCEPT_DB_PATH MARM_DB_PATH MARM_RATE_LIMIT_RPM MARM_SIGNUP_PROMPT_ENABLED RECALL_SCAN_LIMIT SERVER_HOST SERVER_PORT WRITE_QUEUE_ENABLED marm_code_lookup marm_compaction
Topics
agent-memory ai-agents ai-memory claude-code codex-cli developer-tools embeddings indexing-engine knowledge-graph llm local-first long-term-memory mcp mcp-servers model-context-protocol