A specialized skill for chunking, embedding, and ingesting pre-processed text directly into a Qdrant vector database for RAG workflows.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install rag-ingest
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 rag-ingest 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 rag-ingest?
rag-ingest is a dedicated utility within the Openclaw Skills ecosystem designed to handle the final stage of the Retrieval-Augmented Generation (RAG) pipeline. It takes interpreted content from an AI agent and performs the heavy lifting of text chunking and vector embedding before writing the data to a Qdrant collection. This skill ensures that your knowledge base remains up-to-date with searchable, high-dimensional vector data.
By focusing strictly on ingestion, rag-ingest allows developers to build modular workflows where content scraping, summarization, and storage are handled by specialized components. It integrates seamlessly with other Openclaw Skills to create robust AI-driven data pipelines, allowing you to turn raw text into actionable intelligence with minimal configuration.
rag-ingest Use Cases
- Syncing processed research notes into a searchable vector database.
- Automating the ingestion of web articles after summarization by another agent in a pipeline.
- Building a private knowledge base from local documentation or internal notes using Openclaw Skills.
- Feeding structured data into Qdrant for semantic search and AI-driven retrieval applications.
How rag-ingest Works
- The skill receives text content either via command-line arguments or standard input (stdin).
- It validates required parameters such as the document ID and relevant topic tags to ensure data integrity.
- The provided content is split into manageable chunks optimized for embedding models.
- Each chunk is processed through an embedding model using the configured API key.
- The resulting vectors, along with metadata such as source and tags, are written to the specified Qdrant collection.
rag-ingest Setup
To use this skill within the Openclaw Skills framework, you need to configure your environment and call the ingestion script. Ensure you have Node.js installed.
# Set the required environment variables
export QDRANT_URL="your-qdrant-endpoint"
export EMBED_API_KEY="your-embedding-api-key"
# Example: Ingesting content directly via CLI
node skills/rag-ingest/scripts/ingest.mjs \
--doc-id "doc-001" \
--topic-tags "net,security" \
--content "Your pre-processed text content goes here..."
# Example: Ingesting via stdin (piping from another tool)
echo "Content to index" | node skills/rag-ingest/scripts/ingest.mjs \
--doc-id "doc-002" \
--topic-tags "web,http" \
--source "https://example.com/article"
rag-ingest Data Schema & Taxonomy
The rag-ingest skill organizes data within Qdrant using a structured payload to ensure high retrieval accuracy within Openclaw Skills workflows.
| Parameter | Required | Description |
|---|---|---|
--doc-id |
Yes | Unique identifier for the document to prevent duplication or enable updates. |
--topic-tags |
Yes | Comma-separated tags used for filtering and categorical retrieval. |
--content |
No | The actual text to ingest (reads from stdin if not provided). |
--source |
No | The origin identifier (e.g., URL) written to the metadata payload. |
--collection |
No | The Qdrant collection name (defaults to kb_main). |
rag-ingest Advanced Features
- Support for piping output from other Openclaw Skills via stdin for seamless multi-agent automation.
- Automatic document overwriting based on the doc-id, allowing for easy content refreshes and updates.
- Configurable collection targeting to manage multiple distinct knowledge bases within a single Qdrant instance.
- Flexible metadata injection that allows for granular filtering in downstream RAG applications.
SKILL.md
Loading
Related Openclaw Skills
Scrapling Web Scraping Skill
A powerful Python web scraping skill designed to bypass anti-bot protections and extract structured data from complex dynamic websites.
Ceaser Privacy Protocol
A privacy-preserving ETH wrapper on Base L2 using zero-knowledge proofs for shielded transactions and gasless withdrawals.
OpenClaw Model Card
A utility for generating standardized model inventories and high-quality visual model card images from your OpenClaw configuration.
YouTube Video Analyzer (Multimodal)
A sophisticated multimodal analysis tool that synchronizes YouTube transcripts with visual frame extraction to provide comprehensive video understanding.
SeaLegs Marine Forecast
An AI-driven marine weather skill that provides vessel-specific safety classifications and detailed atmospheric data for any maritime coordinate.
KYC Vault
A security-focused automation tool for completing identity verification processes using locally stored documents.
METADATA
Featured*
