A self-contained long-term memory plugin that enables AI agents to automatically capture conversation highlights and recall relevant memories for persistent context.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install install-powermem-memory-full
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 install-powermem-memory-full 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 PowerMem Memory Guide?
The PowerMem plugin is a robust long-term memory solution designed to integrate seamlessly with Openclaw Skills. It solves the problem of context loss in AI conversations by providing a dedicated storage layer that auto-captures valuable insights and intelligently injects them back into the LLM context during future turns.
By utilizing either a local CLI-based SQLite setup or a centralized HTTP server, PowerMem allows developers to build agents that truly learn over time. It leverages the existing Openclaw Skills model configurations to minimize setup friction, ensuring that your agent's memory is powered by the same high-quality LLMs used for primary processing.
PowerMem Memory Guide Use Cases
- Persistent user preference tracking across multiple sessions without manual data entry.
- Building knowledge-base agents that learn from real-time conversation highlights.
- Automating context injection for complex coding or research tasks where previous decisions matter.
- Reducing token overhead by recalling only contextually relevant snippets rather than entire histories.
How PowerMem Memory Guide Works
- Initialization: The plugin checks for Python 3.10+ and verifies the OpenClaw environment for model and provider authentication.
- Auto-Recall: Before a new conversation turn starts, the plugin searches the memory database and injects a relevant-memories block into the prompt context.
- Conversation Processing: The agent interacts with the user, benefiting from the recalled long-term context.
- Auto-Capture: After the interaction, the system analyzes the conversation to extract and save new valuable memories.
- Intelligent Inference: Optionally, the plugin uses an LLM to refine and categorize the captured text for better searchability later.
PowerMem Memory Guide Setup
Ensure OpenClaw is installed and configured with a default model. Then, follow these steps:
# 1. Verify Python 3.10+
python3 --version
# 2. Install PowerMem via pip
python3 -m venv ~/.openclaw/powermem/.venv
source ~/.openclaw/powermem/.venv/bin/activate
pip install powermem
# 3. Install the plugin to your Openclaw Skills environment
openclaw plugins install /path/to/memory-powermem
# 4. Enable and configure via CLI
openclaw config set plugins.enabled true
openclaw config set plugins.slots.memory memory-powermem
openclaw config set plugins.entries.memory-powermem.config.mode cli
openclaw config set plugins.entries.memory-powermem.config.autoCapture true --json
PowerMem Memory Guide Data Schema & Taxonomy
PowerMem organizes data within the OpenClaw state directory. The default structure and metadata taxonomy include:
| Component | Details |
|---|---|
| Storage Engine | SQLite (default) located at <stateDir>/powermem/data/powermem.db. |
| Memory Fields | text, importance, timestamp, and memoryId. |
| Search Mechanism | Vector-based embedding search with configurable scoreThreshold. |
| Configuration | Managed via openclaw config or an optional .env file for fine-tuning. |
PowerMem Memory Guide Advanced Features
- Multi-Mode Support: Switch between local
climode for privacy andhttpmode for shared enterprise environments. - OpenClaw Model Injection: Automatically reuse existing LLM and embedding provider credentials from your Openclaw Skills setup.
- Granular LTM Control: Manual memory management using
ltm add,ltm search, andltm healthcommands. - Custom Inference: Enable
inferOnAddto use AI-driven summarization of memories before they are committed to long-term storage.
SKILL.md
Loading
Related Openclaw Skills
Open Exchange Rate
A zero-config currency conversion and exchange rate skill for AI agents using the public ExchangeRate-API.
Forms for Google Drive
An AI agent skill for seamless Google Forms integration, enabling automated form creation, response analysis, and Excel exports via managed OAuth.
Doc Co-authoring
A highly structured, three-stage workflow for collaboratively drafting, refining, and testing professional documentation with an AI agent.
OpenClaw Workflow Automation
A comprehensive automation suite for building, deploying, and monitoring high-impact AI agent workflows from templates.
Tavily Search
A specialized web search and content extraction tool designed to provide AI agents with clean, high-relevancy data from the live web.
Summarize
A lightning-fast CLI tool designed to generate high-quality summaries of web content, local documents, and multimedia links using advanced LLMs.
Featured*
