A self-hosted long-term memory and cognition layer for OpenClaw agents utilizing 3-signal retrieval and knowledge graphs.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install maasv-memory
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 maasv-memory 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 maasv Memory?
maasv Memory serves as a sophisticated cognition layer designed to replace default memory backends for Openclaw Skills. It provides agents with a structured long-term memory that utilizes 3-signal retrieval—combining semantic search, keyword matching, and knowledge graph traversal—to ensure highly accurate context injection. This system allows agents to move beyond simple chat history into the realm of entity extraction, temporal versioning, and experiential learning.
Built with a focus on privacy and control, maasv is entirely self-hosted. All data is stored within a local SQLite database on your machine, ensuring that no sensitive information is sent to a proprietary cloud service. By integrating this skill, developers can create more intelligent, context-aware agents that maintain a deep understanding of complex relationships and past interactions.
maasv Memory Use Cases
- Building agents that require deep context about user preferences and past project decisions.
- Creating research assistants that maintain a complex knowledge graph of entities and relationships.
- Implementing high-privacy AI workflows where all agent memory must be stored on local hardware.
- Developing agents that perform experiential learning by querying past reasoning logs to improve future outcomes.
How maasv Memory Works
- The maasv server runs on your local machine, hosting a SQLite database and managing embedding generation.
- When an agent interacts with a user, the plugin automatically captures conversation summaries and extracts relevant entities.
- The system generates semantic embeddings and knowledge graph nodes to store the information in a dedup-aware manner.
- Using the autoRecall feature, the skill performs a 3-signal search across the memory store to find the most relevant context for the current prompt.
- The agent uses specific tools like memory_wisdom to record reasoning and outcomes, allowing for future search of past decisions.
maasv Memory Setup
1. Start the maasv Server
First, install the maasv server and its dependencies. You will need to configure your environment variables for your chosen LLM and embedding providers.
pip install "maasv[server,anthropic,voyage]"
cp server.env.example .env
maasv-server
2. Install the Plugin
Add the memory plugin to your OpenClaw environment using the CLI.
openclaw plugins install @maasv/openclaw-memory
3. Activate the Skill
Edit your ~/.openclaw/openclaw.json configuration file to enable the memory slot and point it to your local server.
{
"plugins": {
"slots": { "memory": "memory-maasv" },
"entries": {
"memory-maasv": {
"enabled": true,
"config": {
"serverUrl": "http://127.0.0.1:18790",
"autoRecall": true,
"autoCapture": true,
"enableGraph": true
}
}
}
}
}
maasv Memory Data Schema & Taxonomy
The skill manages data locally via SQLite, organizing information into a multi-layered taxonomy for efficient retrieval in Openclaw Skills.
| Data Type | Description |
|---|---|
| Entities | Extracted subjects, people, or organizations identified during conversations. |
| Knowledge Graph | Mappings of relationships and profiles connecting different entities. |
| Wisdom Store | Logs of agent reasoning and decision outcomes for experiential learning. |
| Embeddings | High-dimensional vector representations for semantic search capabilities. |
| Temporal Data | Versioned records that track how information and memories change over time. |
maasv Memory Advanced Features
- 3-Signal Retrieval: Combines semantic, keyword, and graph-based search for maximum recall accuracy.
- Fully Local Operation: Support for Ollama and Qwen3-Embedding-8B allows for zero data leaving your machine.
- Experiential Learning: Agents can log reasoning and record outcomes to refine future decision-making.
- Entity Extraction: Automatically identifies and categorizes key concepts and relationships from raw text.
- Multi-Agent Support: Provides a shared cognition layer that can be accessed by multiple Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
FFmpeg CLI
A comprehensive media processing toolset for automated video transcoding, editing, and manipulation via the command line.
Hardcover
A GraphQL-based skill for AI agents to query and manage personal reading libraries and book metadata from Hardcover.app.
QMD Memory Skill
A local, high-performance hybrid search and memory engine designed to eliminate API embedding costs for AI agents.
qmd
A high-performance local search engine designed for indexing and retrieving Markdown notes and documents using keyword, vector, and hybrid search methods.
USD1 WLF Transfer
A secure skill for AI agents to perform USD1 (USDC on Wormhole) transfers across wallets using the Wormhole Liquidity Facility.
USD1 WLF Transfer
A specialized skill enabling AI agents to securely transfer USD1 stablecoins across chains using the Wormhole Liquidity Facility.
METADATA
Featured*
