skills.rest
profbernardoj
profbernardojCOMMUNITY
Explore all skills by this author →
GitHub

memory-upgrade

Diagnose and repair OpenClaw memory search with local embeddings and hybrid search.

Updated Feb 18, 2026
Search Tags:#recall#embeddings#memory#local#memory-search#hybrid-search#openclaw
One-click install
npx skills add https://github.com/profbernardoj/travelclaw.org --skill memory-upgrade
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: memory-upgrade
Source: https://github.com/profbernardoj/travelclaw.org/tree/main/memory-upgrade
Command: npx skills add https://github.com/profbernardoj/travelclaw.org --skill memory-upgrade

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires node, openclaw, and includes _scripts_ (resource) components.

What problem does it solve?

OpenClaw memory search is often broken, causing poor cross-session recall. This Skill enables fully local memory search by enabling local embeddings, hybrid BM25+vector search, and session transcript indexing, all running offline without API keys.

Core Features & Use Cases

  • Fully local embeddings and hybrid search for fast recall without external services.
  • Session transcript indexing and an embedding cache to improve cross-session recall.
  • Safe, offline memory upgrades compatible with existing memory files, providing privacy and zero API usage.

Quick Start

Run the quick diagnostic and apply the local-memory patch to restore reliable memory search.

Frequently Asked Questions about memory-upgrade

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix broken OpenClaw memory search when cross-session recall stops working?▼

To fix broken OpenClaw memory search, you can apply a local-memory patch that re-enables embeddings and session transcript indexing, restoring reliable cross-session recall without external services.

How does local hybrid search work for retrieving memory without API keys?▼

Local hybrid search works by combining BM25 text retrieval with local vector embeddings offline. This mechanism indexes session transcripts and caches embeddings to provide fast recall without requiring external API keys.

Can I run fully local embeddings offline with Node and OpenClaw?▼

Yes, you can run fully local embeddings offline with Node and OpenClaw. The memory upgrade enables local embedding generation and hybrid search directly on your device, ensuring privacy and zero API usage.

What is the best way to diagnose missing embeddings in an OpenClaw installation?▼

The best way to diagnose missing embeddings is to run a quick diagnostic check. This detects whether local embeddings and session indexing are active, then applies a patch to verify memory readiness across devices.

Are local memory upgrades compatible with existing OpenClaw memory files?▼

Yes, local memory upgrades are completely compatible with existing memory files. The patch safely enables offline embeddings and automatic file watching without overwriting or corrupting your current memory data.

Related Skills

Explore More Software Engineering Skills→
openclawopenclaw
1.0k

session-viewer

Convert JSONL session transcripts from Codex, Claude Code, OpenClaw, or Pi into interactive HTML files.

Official
Intermediate
EverClawEverClaw
23

memory-upgrade

Diagnose and repair OpenClaw memory search with local embeddings and hybrid retrieval.

Official
Advanced
jason660519jason660519

qmd

Search local notes and transcripts with hybrid BM25, vector, and reranking retrieval.

Community
Intermediate
J4JoshuaJ4Joshua

ollama

Serve local language and embedding models offline on Apple Silicon.

Community
Advanced