Agent♥︎Age
Catalog

open-memex

Official

by stoneskin · TypeScript

Local-first memory for AI coding agents — Markdown + SQLite FTS5, shared across editors via MCP.

io.github.stoneskin/open-memex MCP Server

The MCP server io.github.stoneskin/open-memex provides local-first memory for AI coding agents. It uses Markdown with SQLite FTS5 for search and stores information locally. The resulting memory can be shared across editors via MCP.

🛠️ Key Features

  • Local-first memory for AI coding agents
  • Markdown-based content
  • SQLite FTS5 indexing/search
  • Shared across editors via MCP

🚀 Use Cases

  • Persisting and retrieving coding context for AI coding agents
  • Searching stored notes or memory via FTS5-backed indexing
  • Sharing the same local memory store across multiple editors

⚡ Developer Benefits

  • Structured memory workflow using Markdown
  • Search support through SQLite FTS5
  • Integration point for editor tooling via MCP

⚠️ Limitations

  • Described scope is limited to “local-first memory” with Markdown + SQLite FTS5; no other capabilities are specified.

Topics

ai-agentsai-memoryclaude-codecopilotcursordeveloper-toolsfts5knowledge-managementlocallocal-firstmarkdownmcpmcp-servermodel-context-protocolopencodeopencode-pluginsqlitevscodeskills