A high-performance AI knowledge engine that implements a complete remember, recall, forget, and improve memory cycle with only six lines of code.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install cognee-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 cognee-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 Cognee Memory System?
Cognee Memory System is a sophisticated knowledge engine designed to provide AI agents with long-term, structured memory. Unlike simple vector databases, Cognee utilizes a hybrid approach combining vector embeddings with knowledge graphs to enable semantic and relational reasoning. This ensures that agents using Openclaw Skills can maintain context across sessions, learn from past interactions, and continuously refine their understanding of user preferences and technical data.
By transforming raw documents into a navigable knowledge graph, Cognee allows for automated routing of the most efficient search strategies. Whether you are building a customer support bot or a complex coding assistant, Cognee provides the infrastructure for agents to bridge the gap between temporary chat context and permanent expertise.
Cognee Memory System Use Cases
- Customer Service Agents: Track historical interactions, failed operations, and resolved cases to provide context-aware support.
- SQL Copilots & Knowledge Distillation: Maintain expert SQL patterns, schema structures, and successful workflow models for advanced data analysis.
- Cross-Session Memory: Preserve user preferences and learning progress across multiple independent chat sessions.
- Autonomous Research: Build a persistent knowledge base that evolves as the agent processes new documents and feedback.
How Cognee Memory System Works
- The system ingests data through the remember API, storing information in both vector and graph formats for dual-layer retrieval.
- When a query is made via recall, Cognee automatically routes the request to the optimal search strategy based on the query's intent.
- The improve cycle allows the AI to learn from feedback, updating the knowledge graph to increase future accuracy.
- Outdated or incorrect information is purged using the forget operation to keep the memory clean and relevant.
- Integration with Openclaw Skills enables automatic memory triggers during session starts, tool usage, and user prompts.
Cognee Memory System Setup
To get started with Cognee and integrate it with Openclaw Skills, follow these steps:
Install Cognee
pip install cognee
Install the Openclaw Plugin
npm install @cognee/cognee-openclaw
Configure Environment Variables
# Required: OpenAI API Key
export LLM_API_KEY="your-openai-key"
# Optional: Cognee Cloud Configuration
export COGNEE_SERVICE_URL="https://your-instance.cognee.ai"
export COGNEE_API_KEY="ck_..."
Cognee Memory System Data Schema & Taxonomy
Cognee organizes data using a hybrid taxonomy that balances speed and relational depth.
| Layer | Storage Type | Function |
|---|---|---|
| Session Cache | Volatile/Fast | Immediate context for current conversations |
| Knowledge Graph | Relational | Permanent storage of entities and their relationships |
| Vector Index | Embedding | Semantic similarity search for unstructured data |
Data is managed through datasets (e.g., main_dataset) and can be partitioned by session_id to ensure privacy and organization across different users or agents.
Cognee Memory System Advanced Features
- Hybrid Vector + Graph Search: Combines semantic search with relationship-based discovery for superior context retrieval.
- Automated Lifecycle Hooks: Native integration with Openclaw Skills for PostToolUse and PreCompact memory retention.
- Continuous Learning (Improve API): Programmatically refine stored knowledge based on agent performance and user corrections.
- CLI Visualization: Includes a built-in local UI (
cognee-cli -ui) to explore the generated knowledge graph visually. - Multi-Platform Deployment: Ready for production on Cognee Cloud, Modal (serverless GPU), Railway, or Fly.io.
SKILL.md
Loading
Related Openclaw Skills
Brainstorming
A collaborative AI skill that translates high-level concepts into refined technical designs and documentation before implementation.
TikTok Live Cart Automation
An automated solution for monitoring pinned products in TikTok Live streams and expediting the add-to-cart and checkout process.
GenVR Skills
A professional Node.js CLI toolkit for generating and processing AI images, videos, and 3D panoramas via the GenVR API.
Video Classifier
A powerful browser automation tool for Openclaw Skills that categorizes unstructured long videos through automated tagging and batch processing.
FBS BookWriter
A sophisticated Node-driven writing engine for creating professional books, manuals, and whitepapers through a structured AI-managed workflow.
happy-notes
An intelligent knowledge base assistant for managing documents, performing deep research, and generating structured content like reports and slide decks.
Featured*
