A persistent, file-based memory system that keeps WhatsApp group and direct message contexts isolated, searchable, and loop-free for AI agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install whatsapp-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 whatsapp-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 WhatsApp Memory?
The WhatsApp Memory skill is a specialized context management system designed for Openclaw Skills users who need their AI agents to maintain long-term memory across various WhatsApp conversations. Instead of treating every message as a fresh interaction, this skill enables agents to store and retrieve specific context for different groups and direct messages. By segregating data into sanitized directory structures, it effectively prevents context bleed between different social circles or professional teams.
Built to be incredibly lightweight and cost-effective, this skill uses file-based storage rather than expensive database lookups or high-token LLM summaries. It provides a structured way to track decisions, participant roles, and ongoing tasks, ensuring that your AI agent remains a productive and informed participant in any conversation thread while adhering to strict coordination rules.
WhatsApp Memory Use Cases
- Recalling previous decisions or agreed-upon outcomes in complex project group chats.
- Tracking specific user preferences and pending tasks within direct message threads.
- Managing multiple AI agents in the same group chat to prevent duplicate responses or messaging loops.
- Generating automated weekly digests of all relevant WhatsApp interactions for a quick overview.
How WhatsApp Memory Works
- The agent identifies the conversation type (group or DM) and sanitizes the unique JID or phone number for directory creation.
- A dedicated directory is initialized containing meta.json for identity and various markdown files for different types of memory.
- Before any response is generated, the agent runs a context injection step to read the most recent entries and decisions.
- Post-response, the agent identifies 'worth-logging' information (decisions, tasks, or new people) and appends them to the relevant files.
- Built-in loop prevention logic checks sender IDs and recent session history to ensure the agent does not reply to itself or repeat messages.
WhatsApp Memory Setup
To initialize a new memory context for a chat, use the provided initialization function within your environment:
# Initialize a group chat memory
init_whatsapp_memory "group" "[email protected]" "Project Alpha"
# Initialize a DM memory
init_whatsapp_memory "dm" "+1234567890" "John Doe"
Ensure your agent has the necessary permissions to read and write to the ~/.openclaw/workspace/memory/whatsapp/ directory to maintain the Openclaw Skills data structure.
WhatsApp Memory Data Schema & Taxonomy
The skill organizes data using a hierarchical file structure to ensure isolation and quick retrieval:
| File | Category | Purpose |
|---|---|---|
meta.json |
Metadata | Stores JID/Phone, human-readable name, and creation timestamps. |
context.md |
General | A running log of chronological conversation events and updates. |
decisions.md |
Group Only | Tracks finalized outcomes, approvals, and consensus points. |
people.md |
Group Only | A directory of participants, their roles, and communication styles. |
notes.md |
DM Only | Stores specific tasks, user-defined preferences, and follow-up items. |
WhatsApp Memory Advanced Features
- Global search functionality to query keywords across all historical WhatsApp memory files.
- Multi-PA coordination logic that prevents agents from stepping on each other's toes in shared groups.
- Automated weekly digest generation using shell-based date filtering to summarize recent activity.
- Echo prevention rules that block agents from responding to their own relayed messages or outbound logs.
- Integration-ready design for Git-based backups and morning briefing automations.
SKILL.md
Loading
Related Openclaw Skills
OpenClaw Hook Examples
A comprehensive guide to intercepting and customizing AI agent behavior using the powerful hook system in Openclaw Skills.
Compact Summary: Structured Memory Compression
A structured memory management skill that compresses conversation history into a high-density Markdown format for efficient context retention.
Aliyun Qwen Omni
A powerful multimodal skill for Alibaba Cloud Model Studio that enables simultaneous image, audio, and text processing.
Aliyun Model Studio Crawl and Skill Generation
A dedicated toolset for synchronizing Alibaba Cloud Model Studio documentation with local AI agent skill definitions.
WhatsApp Diagnostics
A comprehensive troubleshooting framework for identifying and resolving WhatsApp connectivity, ingestion, and runtime issues in AI agents.
SDR Project Design
A comprehensive design and planning framework for building modular, service-oriented software-defined radio observatories.
Featured*
