Distributed embedding generation across a device fleet for high-performance RAG and semantic search pipelines.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install fleet-embeddings
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 fleet-embeddings 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 Fleet Embeddings?
Fleet Embeddings is a powerful technical solution designed to distribute text-to-vector transformation tasks across a network of available nodes. By leveraging the Openclaw Skills framework, this tool allows developers to utilize models such as nomic-embed-text, mxbai-embed, and snowflake-arctic-embed without creating bottlenecks on a single machine. The skill ensures that while heavy embedding tasks for RAG pipelines are processed, LLM inference can continue uninterrupted on other devices in the fleet.
This architecture is essential for modern AI applications that require processing thousands of document chunks simultaneously. By routing requests via the Ollama Herd protocol, it provides a scalable way to build knowledge bases, duplicate detection systems, and recommendation engines. Integrating this into your Openclaw Skills setup ensures that your AI agents remain responsive even during heavy data ingestion phases.
Fleet Embeddings Use Cases
- Processing thousands of document chunks for RAG knowledge bases without blocking LLM inference.
- Building large-scale semantic search indexes across distributed hardware setups.
- Implementing high-speed duplicate detection and vector similarity analysis in big data environments.
- Powering recommendation systems that require real-time vectorization of user content.
How Fleet Embeddings Works
- The user starts the central fleet router which listens for embedding requests on a dedicated port.
- Worker nodes are initialized across the network, connecting back to the main router to announce their availability.
- When an embedding request is sent, the router evaluates node health and current load using a 7-signal scoring engine.
- The task is parallelized and dispatched to the optimal node, allowing multiple machines to process batches of text simultaneously.
- The resulting vectors are returned to the application, while performance and usage metrics are logged for monitoring.
Fleet Embeddings Setup
To get started with this skill, install the necessary package and initialize the nodes:
pip install ollama-herd
herd # Start the primary router on port 11435
herd-node # Start the node on each contributing device
ollama pull nomic-embed-text # Ensure the embedding model is available
Fleet Embeddings Data Schema & Taxonomy
The skill manages its operation through local files and structured metadata to ensure reliability within the Openclaw Skills ecosystem:
| File Path | Description |
|---|---|
~/.fleet-manager/latency.db |
Database storing node response times and performance metrics. |
~/.fleet-manager/logs/herd.jsonl |
Structured logs containing request history and routing data. |
metadata |
JSON object for tagging requests (e.g., {"tags": ["indexing"]}) to track per-project usage. |
Fleet Embeddings Advanced Features
- Intelligent load balancing using a 7-signal scoring engine to find the most efficient node.
- Multi-node batch embedding capabilities to significantly reduce the time required for data ingestion.
- Drop-in compatibility with OpenAI SDK and standard HTTP clients for seamless integration.
- Context protection logic to prevent frequent model reloading and minimize latency overhead.
- Built-in analytics dashboard to monitor fleet health and per-app token consumption.
SKILL.md
Loading
Related Openclaw Skills
Coffee Shop Video Maker
A specialized AI video generation tool designed to transform the daily operations of independent cafes and espresso bars into high-converting social media content.
City of Vancouver Open Data
Programmatically access and analyze over 194 City of Vancouver open datasets through a powerful API-driven interface.
ClawStars
ClawStars is a SocialFi layer on the Base blockchain that enables AI agents to trade tickets, share market analysis, and compete in seasonal rewards programs.
Ollama Fleet Router
A smart fleet router that herds multiple Ollama LLM instances into a single, high-performance endpoint with intelligent load balancing.
Local Speech-to-Text Transcription
A secure, fleet-distributed local speech-to-text transcription skill powered by Qwen ASR and MLX for Apple Silicon.
Enterprise Change Management Lead
An expert change management advisor designed to diagnose adoption risks and drive stakeholder readiness for complex enterprise transformations.
Featured*
