🚀 Privacy-First • On-Premises • Edge-Native

Hub-Spoke AI Agent
Orchestration for Edge Mesh

Deploy AI agents across your edge devices — Raspberry Pi, Jetson, phones, laptops — all coordinated through a central hub. Every model runs locally. Your data never leaves your network.

📖 Read the Docs View on GitHub ⚡ Quick Install

∞

Edge Devices Supported

Adapter Types

100%

On-Premises

Data Sent to Cloud

┌─────────────────────────────────────────────────────────────────┐
│                        SeaClip Hub                               │
│  ┌─────────────────┐  ┌─────────────────┐  ┌─────────────────┐  │
│  │  React Dashboard │  │  Express API    │  │  WebSocket Hub  │  │
│  └────────┬────────┘  └────────┬────────┘  └────────┬────────┘  │
│           └────────────────────┼────────────────────┘           │
│  ┌─────────────────────────────┴─────────────────────────────┐  │
│  │  Adapters: SeaClaw · Ollama · Agent Zero · Telegram       │  │
│  └───────────────────────────────────────────────────────────┘  │
└─────────────────────────────────────────────────────────────────┘
              │                │                │
        ┌─────┴─────┐    ┌────┴────┐    ┌─────┴─────┐
        │ 🍓 Pi     │    │ 🎮 Jetson│    │ 💻 Laptop │
        │   Spoke   │    │   Spoke  │    │   Spoke   │
        └───────────┘    └──────────┘    └───────────┘

Built for the Edge

Everything you need to orchestrate AI agents across your on-premises infrastructure.

🌐

Edge Mesh Dashboard

Real-time visibility into all spoke devices — status, telemetry, CPU, RAM, GPU, temperature. Force-directed topology visualization.

→ Monitor your entire fleet at a glance

🦙

Ollama Integration

Route tasks to local LLMs (llama3, mistral, gemma2) running on your devices. GPU utilization tracking and model management.

→ Run AI models without cloud dependencies

🤖

Agent Zero Support

Deep integration with Agent Zero sessions. Memory sync, tool calls, and persistent context across conversations.

→ Powerful autonomous agents on your hardware

📱

Telegram Bridge

Bot-based command interface and push notifications. Control your agents from anywhere via Telegram.

→ Mobile-first agent management

🔗

Multi-Hub Federation

Link multiple SeaClip hubs for cross-site task routing. Sync agents, issues, and telemetry across locations.

→ Scale across data centers and offices

💰

Budget Control

Per-agent cost tracking with configurable spending limits. Monitor token usage and compute costs in real-time.

→ Never exceed your AI budget

🔄

Auto-Reassignment

When a device goes offline, tasks are automatically reassigned to healthy devices. Zero manual intervention.

→ Built-in fault tolerance

⚡

SpacetimeDB Real-time

Live synchronization for Pixel Office visualization and telemetry streaming. See agents move in real-time.

→ Instant updates across all clients

Choose Your Edition

Start simple and scale up as your needs grow.

v1-simple

Base Edition

Core SeaClip functionality for edge mesh orchestration. Perfect for getting started.

Edge mesh dashboard with device monitoring
Ollama adapter for local LLMs
Agent Zero integration
Telegram bridge for notifications
Hub federation for multi-site
Budget control and cost tracking
WebSocket real-time updates
CLI for automation

Install Base

v2-enhanced

Full Feature Set

Everything in Base, plus advanced visualization and workflow features.

Everything in Base edition
9 BMAD agent profiles (Orchestrator, PM, Architect...)
Grammar Shield (prompt injection detection)
Secret filtering (AGENT_ prefix convention)
Vector memory / RAG with embeddings
Pixel Office animated dashboard
GMeet-style multi-agent calls
Git-Agent workflow (job branches, PR tracking)

Install Enhanced

Get Started in Minutes

One-line install scripts for both hub and edge devices.

🏠

Install the Hub

Central server with dashboard and API

bash — One-line install

curl -sSL https://raw.githubusercontent.com/t4tarzan/seaclip/main/scripts/install.sh | bash

Or clone manually:

bash

git clone https://github.com/t4tarzan/seaclip.git
cd seaclip && pnpm install && pnpm dev

📡

Connect a Spoke

Edge device (Pi, Jetson, Linux)

bash — Run on edge device

curl -sSL https://raw.githubusercontent.com/t4tarzan/seaclip/main/scripts/spoke-install.sh | bash -s -- --hub http://YOUR_HUB_IP:51842

Verify connection:

bash

pnpm spoke status

How It Works

Simple setup, powerful orchestration.

Deploy the Hub

Install SeaClip on your central server. The hub runs the dashboard, API, and coordinates all agents.

Connect Spokes

Run the spoke installer on each edge device. They auto-register and start sending telemetry.

Configure Agents

Set up Ollama, Agent Zero, or custom adapters. Assign agents to devices.

Orchestrate

Create tasks, monitor progress, and watch your AI agents work across your edge mesh.

Hub-Spoke AI Agent Orchestration for Edge Mesh