100100 Ideas AI
Submit Tool

MemOS

MCPOpen source

Self-evolving memory OS for LLM & AI agents with hybrid retrieval, multi-modal memory, and 35% token savings.

11kApache-2.0

Self-evolving memory OS for LLM & AI Agents: ultra-persistent memory, hybrid-retrieval, and cross-task skill reuse, with 35.24% token savings and DeepSeek Harness support.

Pros

  • +Unified memory API with graph-structured, inspectable memory
  • +Multi-modal memory supporting text, images, tool traces, and personas
  • +Hybrid retrieval (FTS5 + vector) for accurate recall
  • +Asynchronous ingestion with millisecond latency for production use

Cons

  • −Requires integration with agent frameworks like OpenClaw or Hermes
  • −May have a learning curve for configuring memory cubes and policies
  • −Cloud plugin may incur costs; local plugin requires self-hosting

Target audience: Developers building AI agents or assistants that need persistent, context-aware memory with cross-task skill reuse.

Related tools

Other open-source tools that share tags with MemOS.

valuecell

MCPApache-2.0

Community-driven multi-agent platform for financial applications with AI agents for stock selection, research, and tradi

11k·AI Agents·Open source