Phloem AI Memory Server

Example Use Cases

• 1

  Persistent memory across sessions to maintain context.

• 2

  Semantic search-based memory rather than keyword matching.

• 3

  Citations linked to specific code lines with confidence tracking.

Description

Phloem is a local Model Context Protocol (MCP) server that enables AI tools to maintain persistent memory across sessions by storing semantic memories, citations linked to code, and causal graphs. It works offline with no network requests, supports multiple AI tools through the MCP standard, and automatically adapts memory confidence as code changes. This allows AI agents to have a shared, evolving understanding of project context beyond static text files.