workflowstacks
mcp-server

MemOS: Boost AI Agent Efficiency with Persistent, Self‑Evolving Memory

Founders building LLM‑powered agents get ultra‑persistent memory, hybrid retrieval and cross‑task skill reuse, cutting token usage by over a third. 12k+ GitHub stars.

From the repo · “Self-evolving memory OS for LLM & AI Agents: ultra-persistent memory, hybrid-retrieval, and cross-task skill reuse, with 35.24% token savings”

12,030 stars+155★ this week1,087 forksTypeScriptGuide quality 9/10 · reviewed Oct 8Updated 10/21/2026100% free · open source
What it does

MemOS provides ultra-persistent memory and token savings for founders using large language models (LLMs) and AI agents, enabling efficient and cost-effective operation of these models.

When to use it
  • •When building conversational AI applications that require remembering context across multiple interactions
  • •When optimizing LLMs for token usage to reduce costs and improve performance
  • •When integrating AI agents into workflows that require persistent memory and skill reuse across tasks
Ready-to-paste prompt
Use the following command to test the MemOS API: `node memos_client.ts -m 'What is the capital of France?'`
Heads up: Make sure to set the `HYBRID_RETRIEVAL` environment variable to enable hybrid-retrieval, which is required for ultra-persistent memory and token savings, by running `export HYBRID_RETRIEVAL=true` before starting the MemOS server
Saves to your device
Use with Claude
New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

🛠️ Technical setup

Expect 20–40 minutes in a terminal — or let your AI agent drive it.

Try it instantly — no install
Claude Code
mkdir -p ~/.claude/skills/memos && curl -fsSL https://workflowstacks.com/api/skills/memos/claude-skill -o ~/.claude/skills/memos/SKILL.md
Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)
claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How MemOS: Boost AI Agent Efficiency with Persistent, Self‑Evolving Memory works
Codeflow
Free to inspect

MemOS: Boost AI Agent Efficiency with Persistent, Self‑Evolving Memory is a very large TypeScript project (~293k lines across 1110 code files, plus 453 test files). It is a full software project: use it through its install path rather than reading it end to end. Last commit this month, Apache-2.0 license, has a test suite.

Size
Very large codebase
~293k lines · 1110 code files · days to read — use, don't read
Setup
Developer setup
A real software project. Use it via its install path; don't expect to read it all.
Runs on
Python · Docker
No API keys detected
TypeScript 59%Python 34%HTML 4%JavaScript 2%Shell 1%
Where to start reading
  1. 1
    README.md
    Start here — what it does and how to install it
  2. 2
    AGENTS.md
    The instructions the AI actually follows
  3. 3
    pyproject.toml
    Dependencies and the commands it exposes
  4. 4
    src/__init__.py
    Inside src/ — the main logic begins here
  5. 5
    examples/README.md
    A worked example — copy this to get going
What's in each folder
src/Core code — the actual logic384 files
apps/Sub-packages / plugins1082 files
packages/Sub-packages / plugins55 files
docs/Documentation137 files
examples/Examples you can copy77 files
evaluation/Folder51 files
deploy/Deployment / infrastructure10 files
docker/Deployment / infrastructure7 files
READMEHas testsDocumentedCI checksDocker readyExamples includedApache-2.0 licenseUpdated this month
Quick Actions
Details
Creator
MemTensor
Language
TypeScript
Category
mcp-server
Published
7/6/2025

Are you the creator of this tool? and earn 85% of every sale.

Show it off in your README:

Featured on WorkflowStacks
[![Featured on WorkflowStacks](https://workflowstacks.com/api/badge/memos.svg)](https://workflowstacks.com/skills/memos?utm_source=github&utm_medium=badge)
🔥 Hot this week, in your inbox
Skills like this, every Monday.

The five fastest-growing open-source AI skills, ranked by GitHub star growth. One email a week, unsubscribe anytime.

Do this with a working workflow

Pages we wrote and tested that put a skill like this to work — importable n8n templates, outcome guides and Claude Desktop configs.

MCP config

GitHub MCP config

Search repos, read code, open issues and PRs from Claude.

MCP config

Memory (knowledge graph) MCP config

Persistent memory across chats via a local knowledge graph.

Related skills

More mcp-server tools founders pair with this one.

mcp-server★ 362,094
Rea: Reverse Engineer Apps
Inspect apps down to binary level with Rea. For founders needing code clarity.
mcp-server★ 156,432
open-webui: Self‑hosted AI Chat UI for Faster LLM Integration
Launch a sleek, web‑based AI interface that works with Ollama, OpenAI API, and more—ideal for founders needing a quick, customizable LLM front‑end. 153k+ GitHub stars.
mcp-server★ 132,268
graphify
Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.
mcp-server★ 1,777
relaticle: AI‑powered CRM to streamline sales and team workflows
A self‑hosted Laravel CRM with native AI agents, 30 built‑in MCP tools and a REST API—ideal for founders needing a customizable sales and contact manager. 1.8k+ GitHub stars
mcp-server★ 1,747
Risuai – Build Custom LLM Role‑Play Stories Quickly
Risuai lets founders craft interactive LLM‑driven narratives without coding, perfect for product demos, training bots, or engaging customers. 1.7k+ GitHub stars.
mcp-server★ 3,821
atmosphere
Portable AI agent runtime for the JVM. One @Agent class runs on Spring AI, LangChain4j, Anthropic, or 9 more behind one SPI. Token streaming, tool calls, human approvals, and governance over WebSocket, SSE, gRPC, or WebTransport/HTTP3. Speaks MCP, A2A, and AG-UI.