A persistent knowledge graph memory system that allows AI agents to store and recall facts, decisions, and events using structured PENMAN notation.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install hypabase-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 hypabase-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 Hypabase Memory?
Hypabase Memory provides a sophisticated long-term memory layer for AI agents, transforming raw interactions into a structured, connected knowledge graph. By utilizing Openclaw Skills, developers can enable their agents to remember user preferences, project facts, and complex workflows across different sessions. This skill utilizes the PENMAN notation to categorize information into episodic, semantic, and procedural memory types, ensuring that the agent's context is always relevant and historically accurate.
Built on a robust SQLite backend, it offers a reliable way for agents to build a "mental model" of their environment. Whether it is tracking a project's technical specifications or remembering how a specific user prefers their code to be formatted, this addition to your library of Openclaw Skills provides the persistent context required for truly intelligent automation.
Hypabase Memory Use Cases
- Storing user preferences and long-term decisions to provide personalized agent interactions.
- Tracking project facts, team roles, and specific technical definitions over time.
- Managing task delegations and meeting outcomes within a persistent knowledge base.
- Recording complex procedures and multi-step workflows for future retrieval using Openclaw Skills.
- Linking disparate facts through shared entities to form a comprehensive mental model for the AI.
How Hypabase Memory Works
- The AI agent identifies a significant piece of information during a conversation, such as a preference or a completed task.
- The information is encoded into a PENMAN atom, defining the action (verb) and various roles like subject, object, or locus.
- The remember tool stores this atom in a persistent SQLite database managed via Openclaw Skills.
- When context is needed for a new prompt, the agent uses the recall tool to query the knowledge graph by entity, action, mood, or timeframe.
- Periodic maintenance is performed using the consolidate tool to merge similar entities and strengthen memory associations based on semantic similarity.
Hypabase Memory Setup
Add the MCP server configuration to your local OpenClaw settings file located at ~/.openclaw/openclaw.json. Ensure you have the uv tool installed for automatic dependency management.
{
"mcpServers": {
"hypabase-memory": {
"command": "uvx",
"args": ["--from", "hypabase", "hypabase-memory"],
"env": { "HYPABASE_DB_PATH": "hypabase.db" }
}
}
}
You can also configure the embedder for semantic search by setting the HYPABASE_EMBEDDER environment variable to openai (requires an API key) or fastembed for local processing.
Hypabase Memory Data Schema & Taxonomy
Hypabase Memory organizes information using a role-based taxonomy and specific memory types. Data is stored in a connected graph where entities are linked by actions and verbs. This structure is unique among Openclaw Skills for its ability to represent complex relationships.
| Component | Description |
|---|---|
| Roles | Subject, Object, Instrument, Recipient, Origin, Locus, Attribute, Value |
| Memory Types | Episodic (events), Semantic (facts), Procedural (workflows) |
| Moods | Actual, Planned, Uncertain, Normative, Conditional |
| Persistence | Local SQLite database (.db file) |
The PENMAN format (e.g., (verb :role entity)) ensures that every memory is decomposed into its constituent parts, allowing for highly granular queries.
Hypabase Memory Advanced Features
- Semantic similarity search using customizable embedding models like OpenAI or Sentence-Transformers.
- Automated entity consolidation to prevent duplicate entries and merge synonymous concepts via semantic overlap.
- Temporal filtering for episodic memories using since and before parameters to retrieve specific historical context.
- Support for nested atoms to represent complex beliefs, intentions, or conditional logic within the graph.
- Configurable decay rates for different memory types to optimize the relevance of Openclaw Skills over long durations.
SKILL.md
Loading
Related Openclaw Skills
X Automation Skill
Automate trend research and engagement-driven posting on X through browser-based AI workflows.
LinkedIn Lead Generation & Research
A specialized research skill for identifying and auditing non-technical business founders on LinkedIn for digital service outreach.
LinkedIn Lead Generation & Research
Automate the discovery and verification of non-tech founders on LinkedIn to generate high-quality sales leads for technology services.
Recipe Finder
An AI-powered recipe discovery engine that connects users with meals based on available ingredients and dietary needs.
Cricket Live
A comprehensive cricket tracking suite for OpenClaw agents providing live scores, IPL updates, and automated match alerts.
Podcast Discovery & Generation
An automated end-to-end pipeline to discover trending topics, research them, and generate fact-checked podcast episodes with professional audio.
Featured*
