A lightweight, disk-based memory management system for Openclaw agents that uses Markdown files and keyword search instead of complex embeddings.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memory-lite
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install memory-lite using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Memory Lite?
Memory Lite is a streamlined skill designed for developers who need efficient memory management within Openclaw Skills without the overhead of vector databases or semantic search configuration. By focusing on local Markdown files, it provides a safe, transparent, and low-risk way to store daily logs and long-term insights.
This skill ensures your agent maintains continuity across sessions by writing directly to disk, making it ideal for environments where local-only data persistence is prioritized over complex cloud-based AI retrieval systems. It avoids the need for gateway restarts or configuration changes, making it a plug-and-play solution for enhancing agent context.
Memory Lite Use Cases
- Appending daily activity logs to structured date-based Markdown files for session tracking.
- Storing persistent project insights and long-term facts in a centralized MEMORY.md file.
- Performing fast keyword-based searches across historical agent logs using grep for quick retrieval.
- Generating heuristic-based summaries of recent activities without the need for additional LLM processing costs.
How Memory Lite Works
- The agent or user triggers a memory operation using specialized Python or Bash scripts included in the skill package.
- For daily logging, the skill identifies or creates a file named
memory/YYYY-MM-DD.mdand appends the provided text. - For long-term storage, the skill updates
MEMORY.mdwith durable facts intended for permanent reference. - Search queries are executed via a keyword-based grep script that scans all files within the memory directory.
- Summarization logic uses local heuristics to extract headings and recent bullets, providing a snapshot of recent context.
Memory Lite Setup
To start using this feature within your Openclaw Skills setup, no complex configuration is required. Simply use the scripts directly from your terminal:
# Add a daily note
python3 scripts/memory_add.py --kind daily --text "Your text here"
# Add a long-term memory
python3 scripts/memory_add.py --kind long --text "Durable fact to remember"
# Search memory using keywords
bash scripts/memory_grep.sh "search_term"
# Generate a summary of the last 2 days
python3 scripts/memory_summarize.py --days 2
Memory Lite Data Schema & Taxonomy
Memory Lite organizes information into a predictable directory structure to ensure compatibility with other Openclaw Skills and standard text editors:
| File/Path | Purpose | Format |
|---|---|---|
memory/YYYY-MM-DD.md |
Daily activity logs | Append-only Markdown |
MEMORY.md |
Curated long-term memory | Structured Markdown |
scripts/ |
Utility scripts for add/grep/summarize | Python and Bash |
All files are stored locally on disk, ensuring your data never leaves your environment unless you choose to sync your directory.
Memory Lite Advanced Features
- Zero-config deployment that works immediately without modifying system environment variables.
- Keyword-based search (grep) providing near-instant results regardless of file size.
- Heuristic summary generation that captures context from recent days without external API calls.
- Local-first architecture that treats Markdown files as the immutable source of truth for agent context.
- Safety-focused design that prioritizes appending data over destructive rewrites.
SKILL.md
Loading
Related Openclaw Skills
Neokarma
A persistent personality framework for AI agents to define, evolve, and maintain a consistent identity across sessions.
Prezentit AI Presentation Generator
An AI-powered skill for generating professional presentations and slide decks instantly through natural language commands.
RealmRouter Switch
A zero-friction model manager and switcher for OpenClaw that simplifies API configuration and model selection through a chat-first workflow or CLI.
Garmin Connect Integration
A comprehensive integration that fetches deep physiological metrics from Garmin Connect to provide recovery-aware training nudges and health summaries.
Blinko
A powerful integration for managing Blinko notes via REST API, enabling automated note capture and structured tagging within your AI workflow.
Raindrop.io Bookmark Manager
A comprehensive CLI-driven tool for managing Raindrop.io bookmarks, collections, and tags using Openclaw Skills.
Featured*
