Show HN: MCP Memory – Fast Agent Memory Using Google's OKF and SQLite FTS5

MCP-Memory is a Model Context Protocol (MCP) server that enables AI agents such as Claude Desktop, Cursor, and Windsurf to maintain persistent long-term memory across sessions.

This server records memory using the Open Knowledge Format (OKF v0.2) standard and leverages SQLite FTS5 to support fast key-value lookups and full-text search within milliseconds per second.

Key Features:

• Persistent State Maintenance: Read, store, search, and delete memory snippets even when chat turns or sessions change.
• Dual-Layer Architecture: Automatically syncs all memory to disk as Markdown (.md) files with hierarchical indexes for human browsing, while providing high-performance search via SQLite.
• Namespace Isolation: Manage memory separately by context, such as user settings or project architecture.
• Zero Boilerplate Setup: Quickly configure for various MCP tools via a Python-based setup wizard.

Provided MCP Tools:

• memory_store: Record and update memory in OKF format
• memory_retrieve: Retrieve memory via specific keys and namespaces
• memory_search: Search based on keywords, tags, and namespaces
• memory_get_last / memory_update_last: Manage session checkpoints to help agents immediately identify where their work left off