Skip to content
Kitbay
Submit
MCP server

iflow-mcp_MemoryOS

Runs locallyApache-2.0Updated on Nov 11, 202612,353GitHub stars

This PyPI package, named iflow-mcp_MemoryOS, carries version 1.1.3 of MemOS, the agent memory framework from MemTensor, under a separate package name; its metadata points back to the MemTensor/MemOS repository and documentation.

InstallRepository Checked on Nov 12, 2026

What it does

This PyPI package, named iflow-mcp_MemoryOS, carries version 1.1.3 of MemOS, the agent memory framework from MemTensor, under a separate package name; its metadata points back to the MemTensor/MemOS repository and documentation. MemOS keeps long-term memories for agents in units it calls memory cubes and supports adding, searching, updating and deleting them. The repository code contains an MCP server module that exposes those operations, plus user and cube management, as tools over stdio, HTTP or SSE. The package page and the MemOS README do not explain how to run it as an MCP server, so setup for this package is undocumented.

Capabilities

  • Add, search, read, update and delete memories in a memory cube
  • Create, register, share and dump memory cubes
  • Chat with memory-augmented responses
  • Control the memory scheduler

Read from the project's code; not stated in its documentation. source

Other MCP servers and skills for Memory