A local-first RAG cache that distills documents into structured Markdown and indexes them for hybrid vector and keyword search.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install raglite-local-rag-cache
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 raglite-local-rag-cache 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 RAGLite?
RAGLite is a specialized local-first RAG cache designed to provide your AI agent with a durable, private storage layer for information not included in its original training data. It is particularly effective for managing personal knowledge, school work, medical records, and internal runbooks. As a key component in the Openclaw Skills library, it ensures that sensitive data remains on your local machine while providing high-performance retrieval capabilities.
Unlike traditional knowledge bases, RAGLite focuses on compression-before-embeddings. It distills documents into structured Markdown first, which reduces noise and duplication. This process results in cheaper prompts and more reliable retrieval. By using open-source building blocks like Chroma and ripgrep, it offers a robust alternative to managed vector databases, keeping your data auditable and version-controllable.
RAGLite Use Cases
- Storing and querying local/private knowledge such as personal notes and medical records.
- Managing internal company runbooks and technical documentation securely.
- Reducing LLM prompt costs by distilling documents into essential information before processing.
- Implementing hybrid search strategies that require both semantic vector similarity and exact keyword matching.
- Creating an auditable, human-readable archive of distilled information for version control.
How RAGLite Works
- Condense: The skill processes raw documents and transforms them into structured, low-fluff Markdown outputs using the Openclaw engine.
- Index: These distilled Markdown files are embedded into a local Chroma collection, allowing for organized data management.
- Query: The skill performs a hybrid retrieval by combining vector similarity via Chroma and fast keyword matching via ripgrep.
- Deliver: The most relevant information is retrieved and provided to the agent or user for context-aware responses.
RAGLite Setup
To use this skill within the Openclaw Skills framework, ensure you have Python 3.11+ and ripgrep installed. Then, run the installation script to set up the local virtual environment:
./scripts/install.sh
You must also have a Chroma server running (defaulting to http://127.0.0.1:8100). If your OpenClaw Gateway requires authentication, ensure the OPENCLAW_GATEWAY_TOKEN environment variable is set.
RAGLite Data Schema & Taxonomy
RAGLite organizes its distilled artifacts and metadata to maintain human readability and system efficiency. The structure includes:
| File/Folder | Purpose |
|---|---|
*.tool-summary.md |
Primary distilled summary of the document content. |
*.execution-notes.md |
Metadata and notes related to the distillation process. |
*.outline.md |
Optional structural overview of the source file. |
.raglite/ |
Internal directory for cache, run statistics, and error logs. |
index.md |
A root index file used for navigating the indexed collection. |
nodes/*.md |
Individual document nodes generated when the --nodes flag is used. |
RAGLite Advanced Features
- Hybrid Retrieval: Combines the semantic power of Chroma vector search with the precision of ripgrep keyword search.
- Compression Engine: Uses Openclaw Skills integration to condense large documents into high-signal Markdown before indexing.
- Local-First Privacy: Designed to run entirely on your local network, ensuring sensitive data never leaves your machine.
- Auditable Markdown: Generates human-readable artifacts that can be reviewed, edited, or tracked via Git.
- Scalable Architecture: Start with local storage and easily transition to hosted vector databases if your data needs grow.
SKILL.md
Loading
Related Openclaw Skills
raglite
A local-first RAG cache that distills documents into structured Markdown and indexes them for hybrid vector and keyword retrieval.
DeepthinkLite
DeepthinkLite is a local-first research tool that generates durable markdown artifacts and enforces a time-budgeted workflow for deep thinking.
Anakin Web Data Extraction
A powerful CLI tool for converting websites into clean, structured data using AI-driven scraping and autonomous research agents.
Usage Visualizer
An advanced analytics engine that transforms raw session logs into professional visual reports and text summaries with 100% local processing.
raglite
RAGLite is a local-first RAG cache that distills documents into structured Markdown for efficient hybrid indexing and retrieval.
RAGLite
A local-first RAG cache that distills documents into structured Markdown for high-efficiency hybrid indexing and retrieval.
Featured*
