LangMem vs Memlo: which memory to use
Last updated
LangMem is an open-source Python library from LangChain for adding long-term memory to agents you build, especially with LangGraph: it gives your agent tools to save and search memories during a conversation, a background manager that extracts and consolidates memories, and prompt optimization, all stored in LangGraph's storage layer or your own. Memlo is a hosted memory server for the agents you already use: Claude Code, Codex, claude.ai, ChatGPT and other MCP apps share memories that agents save on purpose, each with its own version history. Pick LangMem to build memory into your own LangGraph agent. Pick Memlo to give your existing agents and your team one shared memory with no code.
At a glance
Facts about LangMem come from its documentation and repository, read on October 31, 2026. Links are in Sources at the end.
| LangMem | Memlo | |
|---|---|---|
| What it is | A library with "tooling to extract important information from conversations, optimize agent behavior through prompt refinement, and maintain long-term memory" | A hosted MCP memory server for your agents |
| Built for | Developers building agents in Python, especially with LangGraph | People and teams who want their agents to share one memory |
| How you use it | pip install -U langmem, then tools such as create_manage_memory_tool and create_search_memory_tool in your agent | A hosted MCP server; nothing to install |
| How memories are made | In the hot path, the agent calls memory tools; in the background, a manager "automatically extracts, consolidates, and updates agent knowledge" | The agent saves a memory with a tool when something is worth keeping |
| Kinds of memory | Semantic (facts), episodic (past experiences) and procedural (system instructions) | Titled notes, a work diary, and a wiki of the people, projects and tools your memories mention |
| Where data lives | "Memory tools use LangGraph's BaseStore for persistence"; it works "with any storage system" | Your Memlo account, hosted in the United States |
| Which agent wrote each change | Not documented | Every version records the person and the connected app or agent that wrote it; memory_history lists them |
| Restore an earlier version | Not documented | Yes: restore_memory brings back any version, as a new version |
| Share with a person | Not part of the library | A space or a single page, by email as a Viewer or an Editor, with a link, or with a team |
| License | MIT | Not open source |
| Price | Free and open source; you pay for your model and storage | Free during the invitation-only beta |
| Can you start today | Yes | By invitation; join the waitlist |
How LangMem works
LangMem offers two ways to manage memory. "Memory management tools that agents can use to record and search information during active conversations 'in the hot path,'" and a "background memory manager that automatically extracts, consolidates, and updates agent knowledge." Its conceptual guide sorts memory into semantic memory ("facts and knowledge"), episodic memory ("past experiences") and procedural memory ("system instructions"), and the library can refine an agent's prompt from what it learns.
It provides "functional primitives you can use with any storage system and native integration with LangGraph's storage layer," and its memory tools "use LangGraph's BaseStore for persistence."
How Memlo works
Memlo is an MCP server at https://memlo.ai/mcp. You add it to each agent and sign in once, with no API key. Agents get tools to save, search, read, change and retire memories.
- One memory for agents from different vendors. Claude Code, Codex, claude.ai and ChatGPT find the same memories when they sign in to the same account.
- History, with authors. Every change is a version, with when it was written and by whom. Any version can be restored, and
edit_memoryrefuses a write when someone else changed the note first. - Spaces and sharing. Keep memories in spaces, and share a space or a page with people by email as a Viewer or an Editor.
Where LangMem is the better choice
- You're building a LangGraph agent. LangMem plugs into LangGraph's store and your agent's tools. Memlo has no SDK.
- You want background extraction. LangMem's memory manager pulls memories out of conversations on its own.
- You want prompt optimization. LangMem can refine an agent's instructions from experience.
- You control the storage. Use LangGraph's store or your own database. Memlo is hosted only.
- You want to start today. It's MIT open source; Memlo is by invitation.
Where Memlo is the better choice
- You want memory for agents you didn't build. Add one URL to Claude Code, Codex, claude.ai or ChatGPT and sign in.
- You use agents from more than one vendor. They all read and write the same memories.
- You want to read and correct each memory. Memories are titled notes you can open, edit and roll back in the Memlo app.
- You want to see which agent wrote what. Every version names the person and the connected app or agent.
- You share with people. Share a space or a single page by email with a role.
What Memlo doesn't do yet
- Cursor. Memlo's sign-in doesn't work from Cursor yet.
- Automatic capture. Agents save memories on purpose; nothing extracts memories from conversations in the background.
- An SDK or API keys. Memlo is reached over MCP with your own sign-in, while LangMem is a library for your code.
- Self-hosting. Memlo is hosted only.
- Open sign-up. Accounts are by invitation for now; you can join the waitlist.
Frequently asked questions
What is LangMem?
An open-source Python library from LangChain that helps agents "learn and adapt from their interactions over time," with memory tools, a background memory manager and prompt optimization, integrated with LangGraph's storage.
Is LangMem open source?
Yes, under the MIT license.
Does LangMem need LangGraph?
It integrates natively with LangGraph's storage layer, and its documentation says its primitives work "with any storage system."
Is there a LangMem alternative that's hosted?
Memlo is one if you want memory for your existing agents rather than a library: Claude Code, Codex and chat apps connect to the same server. For developer frameworks, see Letta vs Memlo and Mem0 vs Memlo.
Sources
Checked on October 31, 2026:
- LangMem documentation and its conceptual guide
- LangMem on GitHub and its license
- Memlo: the setup page, how it works and terms
Try Memlo
Memlo is invitation only for now. Join the waitlist and we'll email you when your invitation is ready.
Join the waitlist