June 6, 2026
Introducing CMEM Cloud: your agent's memory, linked everywhere
For the last year, claude-mem has done one stubborn, useful thing: while one AI works, a second AI takes notes. The result is a memory that persists across sessions — your agent remembers what it did, what you decided, and why, searchable across every project, stored local-first on your own machine.
That foundation is open source, and it stays open source. Today we're opening early access to the next layer: CMEM Cloud.
What CMEM Cloud is
CMEM Cloud is the cloud that links your local-first memory everywhere. One private MCP link gives any MCP client — Cursor, Codex CLI, Gemini CLI, and more — secure access to your memory from anywhere, without giving up the local-first foundation that keeps your data yours.
You keep the engine you already trust. The cloud just makes it reachable from every tool you work in, instead of only the machine where the notes were taken.
Why we built it
The memory was always good. Getting to it from a second editor, a second machine, or a teammate's setup was the hard part. A private MCP link solves that directly: it's one endpoint, scoped to you, that any MCP-speaking client can connect to. No re-indexing per tool. No re-explaining your context every time you switch editors.
The engine underneath
CMEM Cloud is built on the same claude-mem internals that ship today. A few things worth knowing about how that engine performs:
- Structural codebase exploration runs 6–12× cheaper than reading files whole, and file understanding lands around 8× cheaper than full reads.
- Model tier routing cuts agent cost by ~52% in production-measured runs.
- The session index stays light — around 800 tokens for 50 observations.
Every one of those figures comes from the project's own measurements; we don't quote numbers we can't point at.
Who it's for
If you already run claude-mem and want that memory available in more than one place — across editors, across machines, across a team — CMEM Cloud is for you. The eleven bundled skills, the six chat integrations, and the structural search you rely on all come along; the cloud just removes the "but only here" caveat.
Getting early access
CMEM Cloud is in early access now. The engine is free and open source; the cloud is the part we're inviting people into a few at a time so we can keep the experience tight.
If that resonates, join the waitlist and we'll reach out as spots open.