A sophisticated three-tier memory architecture providing sliding windows, automated summarization, and vector-based long-term storage for AI agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install monica-memory-manager
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 monica-memory-manager 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 Manager?
The Memory Manager is a robust framework designed to solve the context window limitations of modern LLMs. By integrating this into your Openclaw Skills workflow, you can provide AI agents with a hierarchical memory structure: a sliding window for immediate interactions, an automated summary layer for medium-term history, and a vector-based Retrieval-Augmented Generation (RAG) system for permanent knowledge retention. This ensures that agents maintain high performance and relevance without losing critical data during extended sessions.
Built for developers who need persistent agent identity and context, this skill optimizes token usage by offloading older information to compressed summaries or searchable vector databases. Whether you are building a personal assistant or a complex coding agent, managing memory effectively through these Openclaw Skills is essential for building stateful, reliable AI applications.
Memory Manager Use Cases
- Managing long-form dialogue history without exceeding token limits.
- Maintaining user preferences and personal knowledge bases across multiple sessions.
- Optimizing API costs by using cheaper models for background summarization tasks.
- Injecting relevant historical context into new conversations via semantic search.
How Memory Manager Works
- Initialize the memory environment to create the necessary directory structure and configuration files.
- Record real-time interactions in a short-term FIFO (First-In-First-Out) sliding window to maintain immediate coherence.
- Monitor token counts and trigger automated summarization using a lightweight model once the defined threshold is reached.
- Archive summaries and important data into a local ChromaDB vector store for permanent storage.
- Perform semantic queries against the long-term memory store at the start of new sessions to retrieve and inject relevant context.
Memory Manager Setup
First, initialize the memory system using the provided script:
python3 scripts/memory_manager.py init
Configure your memory parameters in memory/config.yaml to define window sizes and thresholds. You can then add data manually or via automation:
# Add specific context to short-term memory
python3 scripts/memory_manager.py add --type short --content "User prefers Python for backend tasks"
# Query existing knowledge
python3 scripts/memory_manager.py search "coding preferences"
Memory Manager Data Schema & Taxonomy
The skill organizes data into three distinct layers to ensure efficient retrieval within the Openclaw Skills ecosystem:
| Tier | Storage Path | Mechanism | Primary Use |
|---|---|---|---|
| Short-Term | memory/sliding-window.json |
JSON FIFO | Immediate conversation flow |
| Medium-Term | memory/summaries/ |
Daily JSON Files | Context compression and history |
| Long-Term | memory/vector-store/ |
ChromaDB | Semantic retrieval and RAG |
Memory Manager Advanced Features
- Configurable sliding window size and max token limits for fine-grained context control.
- Multi-model support for summarization, allowing the use of efficient models like GLM-4-Flash or GPT-3.5.
- Semantic relevance filtering with adjustable similarity thresholds (min_relevance).
- Local vector database support via ChromaDB for privacy and speed without external dependencies.
SKILL.md
Loading
Related Openclaw Skills
Video Ad Deconstructor
An AI-powered analysis tool that deconstructs video ad creatives into actionable marketing insights like hooks, social proof, and emotional triggers.
Video Ad Analyzer
An AI-driven toolkit for extracting text, scenes, and transcriptions from video advertisements using Google Gemini Vision.
Meta Ad Creatives
A data-driven performance tracking skill for analyzing Meta Ads creative hit rates and cross-account benchmarks.
YouTube Idea Extractor
A specialized tool for mining YouTube transcripts to identify and validate high-potential startup opportunities and business ideas.
Memory Manager
A sophisticated three-tier memory management system for AI agents that handles short-term context, mid-term summaries, and long-term vector storage.
OpenClaw Auto-Doctor
An intelligent maintenance assistant that monitors logs to automatically diagnose errors, search for solutions, and apply fixes or create pull requests.
Featured*
