Persistent Claude Code memory, but with a real runtime around it

When people search for Claude Code memory, they usually want more than saved snippets. They want the assistant to keep operator context across sessions, remember past decisions, avoid repeated mistakes, and stay local. That is where NEXO fits best today.

Deepest current fit Local shared brain Guard + doctor + workflows
Terminal
$ npx nexo-brain
$ nexo remember "Release checklist changed"
$ nexo recall "release checklist"
1

Memory that survives session resets

NEXO persists memory locally instead of treating the current context window as the whole world. That matters when work spans many terminal sessions.

2

Guard before action

The runtime can surface relevant learnings and blocking rules before edits, which is more useful than simple retrieval after the mistake already happened.

3

Operational surfaces included

Claude Code gets more than memory: doctor diagnostics, durable workflows, personal scripts, followups, and review-gated Evolution all live in the same runtime.

It is strong for Claude Code today

Claude Code remains the recommended path because it has the deepest shared-brain integration and the most complete operator workflow surface.

It is still libre/open-source software

The full source is inspectable, modifiable, and self-hosted on your own machine. There is no hidden hosted memory plane behind the integration.

It is not only a memory plugin

NEXO is a local cognitive runtime. Memory is central, but the surrounding reliability and workflow surfaces are part of the value.

Want the exact install and parity path?

The docs go deeper on install flow, client parity, and how Claude Code shares the same brain with Codex and Claude Desktop.