Skip to content

Latest commit

 

History

History
69 lines (42 loc) · 3.06 KB

File metadata and controls

69 lines (42 loc) · 3.06 KB

OpenClaw Migration Guide

ClawVault is deprecated as the recommended memory path for OpenClaw.

OpenClaw now has a first-party memory stack with official docs, builtin memory, and a maintained QMD backend. For new deployments, use OpenClaw native memory instead of installing ClawVault as a plugin.

Why this changed

ClawVault got to this shape early.

It helped prove that markdown-native structured memory could work well in practice, especially with:

  • local, human-readable memory files
  • Obsidian-friendly workflows
  • QMD-style local retrieval for stronger recall

Those ideas are now part of OpenClaw's official, maintained memory story. OpenClaw's memory docs and QMD backend are the path that should move forward.

Recommendation

New deployments

Choose OpenClaw native memory.

Do not start a new OpenClaw install by wiring in ClawVault unless you are deliberately maintaining an older setup.

Existing ClawVault users

If your current setup works, you do not need to panic-migrate immediately. But you should treat ClawVault as legacy integration and plan to move onto OpenClaw's first-party memory stack.

Migration direction

The usual mapping is straightforward:

ClawVault concept OpenClaw path
ClawVault plugin as memory layer OpenClaw builtin memory or QMD backend
Vault-backed memory workflow MEMORY.md, memory/YYYY-MM-DD.md, and OpenClaw memory tooling
QMD-style retrieval via ClawVault-era setup OpenClaw's maintained QMD Memory Engine
OpenClaw memory entrypoint questions Memory Overview

What to read now

Start with the official OpenClaw docs:

  1. Memory Overview
  2. Builtin Memory Engine
  3. QMD Memory Engine

Use builtin memory when you want the default, supported path with no extra sidecar.

Use QMD when you want local-first retrieval with reranking, query expansion, transcript indexing, or extra indexed paths.

Practical migration notes

  • Prefer OpenClaw's native memory files and tools as the source of truth going forward.
  • Move any new setup or documentation toward OpenClaw memory concepts, not ClawVault plugin registration.
  • If you relied on ClawVault mainly for markdown-native memory plus local retrieval, OpenClaw now covers that directly.
  • If you relied on Obsidian-friendly file workflows, keep the file-oriented workflow, but migrate the runtime integration to OpenClaw's memory stack.

Historical note

ClawVault was useful because it arrived before the official path existed. It helped establish the shape of markdown-native structured memory in the OpenClaw ecosystem.

That is worth preserving. It is also why this repository now points new users to OpenClaw first.