A cloud-native persistent memory skill that enables OpenClaw agents to automatically capture facts and recall context across conversations.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install seekdb-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 seekdb-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 SeekDB Memory?
SeekDB Memory provides a sophisticated long-term memory layer for agents using Openclaw Skills. It moves beyond session-based context by offering cloud-native persistence, ensuring that your agent remembers user preferences, past decisions, and key project details regardless of the device or session. By utilizing a hybrid search approach that combines vector similarity with fulltext search and reranking, it ensures the most relevant information is always at the agent's fingertips.
This skill is essential for developers building complex assistant workflows where context continuity is critical. It automates the tedious parts of memory management—like fact extraction and context injection—allowing the agent to behave more intelligently and personally over time.
SeekDB Memory Use Cases
- Recalling user preferences across different chat sessions to provide personalized recommendations.
- Maintaining a history of technical decisions and project requirements during long-term development cycles.
- Automatically capturing and organizing deadlines or important facts shared during a conversation.
- Retrieving specific past interactions when a user references previous discussions or "last time."
How SeekDB Memory Works
- The plugin automatically performs a hybrid search using vector similarity and fulltext search before each reply to find relevant memories based on the user's message.
- Relevant memories are injected directly into the agent's prompt context, categorized as relevant-memories.
- After the conversation concludes, the system uses an LLM to extract key facts, decisions, or preferences from the dialogue.
- These extracted facts are stored in the cloud with specific metadata like importance scores and categories for high-precision future retrieval.
- The system tracks successful tool-use patterns and distills them into reusable experiences to optimize future task performance.
SeekDB Memory Setup
To get started with this memory layer for Openclaw Skills, run the automated setup command:
/openclaw-m0-setup
Alternatively, install the m0 plugin manually via the CLI:
openclaw add m0
Finally, configure your API key and preferences in your ~/.openclaw/openclaw.json file to enable auto-capture and auto-recall functionality.
SeekDB Memory Data Schema & Taxonomy
SeekDB Memory organizes data into facts and experiences with a structured metadata taxonomy to ensure high-precision retrieval within Openclaw Skills.
| Attribute | Description |
|---|---|
| text | The core content of the memory or fact. |
| metadata.category | Classification such as preference, decision, fact, or deadline. |
| metadata.importance | A numerical score from 0.0 to 1.0 indicating the weight of the memory. |
| experience_id | Linkage to specific distilled successful patterns or tool-use recipes. |
SeekDB Memory Advanced Features
- Hybrid Search Engine: Combines HNSW vector similarity with BM25 fulltext search and RRF fusion for superior context accuracy.
- Intelligent Query Rewriting: Resolves ambiguous pronouns and expands keywords to improve memory retrieval success.
- Experience Learning System: Automatically distills successful complex workflows into reusable patterns to avoid repeating past mistakes.
- Cross-Device Synchronization: Cloud-native storage ensures that memories are consistent across any device or interface running the agent.
SKILL.md
Loading
Related Openclaw Skills
Agent Reach
A comprehensive connectivity toolkit for AI agents to search, read, and interact across 13+ major social and professional platforms.
PDF to Markdown Converter
A professional-grade conversion tool that utilizes the MinerU API to transform PDF documents into structured, clean Markdown files.
Netfluid
Netfluid provides the payment rails for AI agents to manage fiat and crypto transactions with full regulatory compliance.
Pipeworx Books
A powerful book discovery and metadata retrieval skill powered by the Open Library catalog.
Kalshi F1 Constructor Trader
An automated trading skill that utilizes constructor performance ratings to execute data-driven F1 championship trades on Kalshi.
Toolfolio
Toolfolio is an automated video analysis tool that detects shot transitions and identifies 14 types of camera movements to generate AI video prompts.
Featured*
