A robust SQLite-backed storage system for AI agents that provides long-term memory, metadata management, and semantic retrieval.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-memory-persistence
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 agent-memory-persistence 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 Agent Memory Persistence?
Agent Memory Persistence is a sophisticated tool designed to give AI agents a durable, long-term memory. By utilizing SQLite for storage and TypeScript for vector processing, it allows developers to build agents that remember user interactions and relevant data across multiple sessions. This addition to the Openclaw Skills ecosystem ensures that agents maintain context and continuity without the overhead of complex external vector databases.
The skill provides a structured way to store text, metadata, and vector embeddings, making it essential for building sophisticated RAG-based workflows. Since it handles cosine-similarity search natively in TypeScript, it offers a high degree of portability for various deployment environments where simplicity and performance are key.
Agent Memory Persistence Use Cases
- Maintaining stateful context for AI agents across different user sessions.
- Building knowledge bases for autonomous agents to reference during long-running tasks.
- Implementing user-specific memory constraints and privacy-focused data expiration.
- Performing semantic search to retrieve relevant past interactions based on vector similarity.
How Agent Memory Persistence Works
- The developer instantiates a MemoryManager by pointing it to a SQLite database file.
- Memories are ingested into the system with associated content, metadata tags, and vector embeddings.
- The VectorIndex component manages similarity search logic to find related context based on query embeddings.
- Agents query the store using specific filters such as session IDs, time windows, or user identifiers.
- The system performs periodic maintenance through lifecycle operations like cleanupExpired to manage database size.
Agent Memory Persistence Setup
To integrate this into your project, ensure you have SQLite installed. Then, initialize the memory manager as follows:
npm install agent-memory-persistence
Configure the storage path in your application:
import { MemoryManager } from './src/MemoryManager';
const memory = new MemoryManager('path/to/memory.sqlite');
Agent Memory Persistence Data Schema & Taxonomy
The skill organizes memory data into a structured SQLite format, enabling efficient lookups and metadata filtering as part of the Openclaw Skills framework:
| Field | Type | Description |
|---|---|---|
| content | String | The raw text of the memory entry. |
| metadata | JSON | Structured data for filtering (user, session, etc.). |
| embedding | JSON Array | Vector representation for semantic search. |
| timestamp | Date | When the memory was created. |
| expiry | Date | Optional TTL for the memory item. |
Agent Memory Persistence Advanced Features
- Semantic retrieval using a TypeScript-native cosine similarity implementation for easy deployment.
- Built-in memory lifecycle management for automated data pruning and stale memory cleanup.
- Multi-tenant support through structured user, session, and type-based filtering.
- Extensible architecture designed to allow swappable ANN indices or SQLite vector extensions for high-volume use cases.
SKILL.md
Loading
Related Openclaw Skills
Agent Communication Hub
A high-performance communication layer providing direct messaging, pub/sub events, and persistent storage for autonomous agent interactions.
Feishu Image Message Skill
A comprehensive toolkit for automating image uploads, messaging, and visual analysis within the Feishu (Lark) communication platform.
Feishu FAQ Bot
A high-performance FAQ management tool for Feishu bots that builds searchable knowledge bases from Markdown files using fuzzy matching.
tmux-terminal
A powerful terminal automation skill that leverages tmux to drive interactive TUI applications and manage persistent background processes.
Agent Task Queue
A high-performance TypeScript task queue for OpenClaw that orchestrates multi-agent workflows with priority scheduling and dependency management.
Agent Usage Tracker
A TypeScript-powered SQLite utility for real-time tracking of AI model token usage and automated budget enforcement.
Featured*
