A local RAG system that provides AI agents with persistent memory and semantic search capabilities across past sessions and notes.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ouyang
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 ouyang 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 Jasper Recall?
Jasper Recall is a robust local Retrieval-Augmented Generation (RAG) system engineered to give AI agents a functional memory. By utilizing ChromaDB and sentence-transformers, it enables agents to index, store, and retrieve information from past conversations, session logs, and markdown notes. This ensures that agents can maintain context over long periods, making it an essential component for developers building sophisticated agents with Openclaw Skills.
The system runs entirely locally, using the all-MiniLM-L6-v2 model to generate 384-dimensional embeddings. This architecture provides high-performance semantic search without the need for external API calls, ensuring data privacy and reducing latency during agent interactions.
Jasper Recall Use Cases
- Persistent context retrieval for AI agents across multiple sessions.
- Semantic search over daily developer logs and project documentation.
- Automated creation of a searchable knowledge base from agent activity.
- Maintaining continuity in long-term autonomous agent workflows.
How Jasper Recall Works
- The digest-sessions tool processes raw agent logs to extract key topics and tool usage data.
- The index-digests command chunks markdown files and generates vector embeddings using a local transformer model.
- Embeddings and metadata are stored in a local ChromaDB instance located in the user's home directory.
- The recall command performs semantic similarity searches against the vector database to return relevant context snippets.
Jasper Recall Setup
To get started with this addition to your Openclaw Skills, run the automated setup command:
npx jasper-recall setup
This command initializes a Python virtual environment, sets up the ChromaDB database, and installs the necessary CLI scripts in your local path.
Jasper Recall Data Schema & Taxonomy
Jasper Recall organizes its data within the ~/.openclaw/ directory. The schema for indexed content includes:
| Component | Path | Description |
|---|---|---|
| Memory Files | ~/.openclaw/workspace/memory/*.md |
Daily notes and core memory files |
| Session Digests | ~/.openclaw/workspace/memory/session-digests/ |
Summaries of past agent interactions |
| Repo Docs | ~/.openclaw/workspace/memory/repos/ |
Project-specific documentation indexed for RAG |
| Vector DB | ~/.openclaw/chroma-db |
The ChromaDB instance holding vector embeddings |
The indexing process uses content hashing to ensure only modified files are updated in the database.
Jasper Recall Advanced Features
- Smart chunking with 500-character windows and 100-character overlap for optimal context preservation.
- Support for automated heartbeats and cron-scheduled re-indexing to keep memory fresh.
- Similarity score visibility via verbose mode for debugging retrieval accuracy.
- JSON output mode for programmatic integration into agent tool-calling logic.
SKILL.md
Loading
Related Openclaw Skills
Internal Comms Skill
A specialized skill for generating high-quality internal corporate communications using standardized company formats and templates.
Deep Research Agent
A sophisticated AI agent that decomposes complex questions into structured research plans and synthesizes data-driven insights.
Deep Research Agent
A sophisticated AI agent that decomposes complex research goals into structured plans and synthesized reports.
Deep Research Agent
A sophisticated AI agent that decomposes complex queries into structured research plans to deliver synthesized, data-driven insights.
MailChannels Email API Integration
A robust integration for sending emails and processing real-time delivery webhooks via MailChannels.
Stock Research Engine
An AI-powered equity research tool that generates institutional-grade fundamental analysis for global stocks from a buy-side perspective.
Featured*
