An automated memory preservation skill that periodically distills agent sessions into daily logs and long-term knowledge bases.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install heartbeat-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 heartbeat-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 Heartbeat Memory?
Heartbeat Memory is a robust session management and knowledge retention tool designed for AI agents. It addresses the challenge of ephemeral chat histories by automatically scanning, summarizing, and archiving interactions into a permanent format. By integrating deeply with the Openclaw Skills ecosystem, it ensures that project decisions, technical discoveries, and context are preserved across sessions without manual intervention.
The skill operates in the background, utilizing a heartbeat mechanism to maintain a continuous record of activity. It doesn't just save text; it uses Large Language Models to distill raw logs into structured daily notes and refined long-term memory files, making it an essential component for developers and researchers who rely on long-form AI collaboration.
Heartbeat Memory Use Cases
- Automatically documenting architectural decisions made during complex AI coding sessions.
- Creating a searchable daily log of AI-assisted research and development activities.
- Synchronizing long-term context across multiple workspaces to keep agents informed of past progress.
- Backing up session data before manual deletion to ensure no critical insight is lost.
How Heartbeat Memory Works
- The skill is triggered automatically every 30 minutes via the Openclaw heartbeat or manually by user command.
- It scans the filesystem to identify new or updated chat sessions that haven't been processed yet.
- The skill fetches session histories and intelligently adjusts the message volume for optimal LLM distillation.
- A sub-agent is spawned to generate structured Markdown daily notes, including tags, summaries, and key discovery points.
- On a weekly schedule, the skill aggregates daily notes into a centralized long-term memory file.
- It automatically updates the workspace status and configuration hashes to track progress and prevent duplicate processing.
Heartbeat Memory Setup
Prerequisites
Ensure you have the Openclaw gateway running and the CLI tools installed.
Installation
Choose one of the following methods:
# Recommended: Install via ClawHub
clawhub install heartbeat-memory
# Alternative: Manual clone
git clone https://github.com/JustZeroX/skill-heartbeat-memory.git ~/.openclaw/skills/heartbeat-memory
Configuration
You must enable the heartbeat feature in your global configuration:
# Set the heartbeat interval to 30 minutes
openclaw config set agents.defaults.heartbeat '{"every": "30m"}'
# Restart the gateway to apply changes
openclaw gateway restart
Heartbeat Memory Data Schema & Taxonomy
The skill organizes data within the ./memory/ directory relative to the current workspace root. It utilizes the following taxonomy:
| Path | Purpose | Format |
|---|---|---|
./memory/daily/YYYY-MM-DD.md |
Daily session summaries and activity logs. | Markdown |
./MEMORY.md |
Long-term distilled knowledge and persistent context. | Markdown |
./memory/heartbeat-memory-config.json |
Workspace-specific processing rules and date filters. | JSON |
./memory/heartbeat-state.json |
Internal state tracking and processing checkpoints. | JSON |
./HEARTBEAT.md |
Automatic status reporting of memory sync activities. | Markdown |
Heartbeat Memory Advanced Features
- Intelligent incremental updates that only process new content in active sessions to save tokens.
- Support for multi-workspace environments with localized configuration detection.
- Automated sub-agent spawning for parallel session distillation and high-volume processing.
- Configurable refinement schedules for long-term memory (e.g., custom weekly intervals).
- Automatic HEARTBEAT.md maintenance with hash-based change detection to avoid redundant file writes.
SKILL.md
Loading
Related Openclaw Skills
OpenViking Pro
A cloud-based memory compression API that reduces AI token costs by up to 96% without requiring any local installation.
Memory-latch
A deterministic recovery primitive designed to restore agent continuity and user trust after context resets or memory outages.
LobstrHunt
LobstrHunt connects your agent to a daily skill launch platform to scout new capabilities and contribute to the community ecosystem.
Pine Script Developer
A specialized tool for writing, debugging, and optimizing TradingView Pine Script indicators and automated trading strategies.
ClawSpa
ClawSpa is a comprehensive wellness and maintenance suite designed to optimize AI agent performance through memory cleanup, security audits, and alignment diagnostics.
De-AI Writing (Chinese)
A specialized skill designed to strip away the robotic, formulaic nature of AI-generated Chinese text to produce natural, professional, and high-density content.
Featured*
