Securely bridge remote AI agents to your local machine
Features • Quick Start • Installation • User Guide • Contributing
简体中文 | English
OpenClaw agents run on your server, but often need to interact with your local environment — running commands, controlling browsers, accessing local files.
OpenClaw Connector solves two problems:
-
Bridging server and local — It creates a secure tunnel between OpenClaw agents and your local machine, so agents can execute tasks on your computer as if they were sitting next to you.
-
Secure gateway access — The OpenClaw gateway should never be exposed to the public internet. This app lets you access it securely from your local machine through an encrypted SSH tunnel — no open ports, no public URLs.
- SSH Tunnel — Secure reverse tunnel to your Linux gateway with auto-reconnect
- Agent Bindings — Bind AI agents to your local node for remote task execution
- Browser CDP — Expose local Chrome browser to agents via Chrome DevTools Protocol
- Session Management — Notify agents across chat sessions with one click
- Device Identity — Ed25519 keypair for secure device authentication
- Emergency Disconnect — One-click kill switch to instantly sever all connections
Download the latest release for your platform from the Releases page.
| Platform | File |
|---|---|
| macOS (Apple Silicon) | OpenClaw.Connector_x.x.x_aarch64.dmg |
| macOS (Intel) | OpenClaw.Connector_x.x.x_x64.dmg |
| Linux (Debian/Ubuntu) | OpenClaw.Connector_x.x.x_amd64.deb |
| Linux (AppImage) | OpenClaw.Connector_x.x.x_amd64.AppImage |
| Windows | OpenClaw.Connector_x.x.x_x64-setup.exe |
macOS note: On first launch, macOS may show "cannot verify developer". Go to System Settings → Privacy & Security, scroll down and click "Open Anyway".
Linux AppImage: Run
chmod +x OpenClaw.Connector_*.AppImagefirst.
- A running OpenClaw gateway on a remote Linux server
- SSH access to the server from your local machine
New to OpenClaw? See the User Guide for detailed setup instructions and parameter explanations.
# Clone the repo
git clone https://github.com/liuzeming-yuxi/Openclaw-Connector.git
cd Openclaw-Connector
# Install dependencies (requires Node.js 18+, pnpm, Rust toolchain)
pnpm install
# Run in development mode
pnpm tauri dev
# Or build for production
pnpm tauri build| Layer | Technology |
|---|---|
| Desktop Framework | Tauri 2 |
| Frontend | React 19 + TypeScript + Tailwind CSS 4 |
| State Management | Zustand 5 |
| Backend | Rust (Tokio async runtime) |
| Tunnel | SSH reverse port forwarding |
| Browser Automation | Chrome DevTools Protocol (CDP) |
├── src/ # React frontend
│ ├── pages/ # Page components
│ ├── components/ui/ # Reusable UI components
│ ├── store/ # Zustand state stores
│ └── types/ # TypeScript type definitions
├── src-tauri/ # Rust backend
│ └── src/
│ ├── lib.rs # Tauri command handlers
│ ├── ssh_tunnel.rs # SSH tunnel management
│ ├── browser.rs # Chrome CDP lifecycle
│ ├── ws_client.rs # WebSocket client
│ ├── config.rs # Configuration persistence
│ ├── health.rs # Gateway health monitoring
│ └── device_identity.rs # Ed25519 device keys
├── docs/ # Documentation
└── package.json
# Run frontend + backend in dev mode
pnpm tauri dev
# Run frontend tests
pnpm test
# Run Rust tests
cargo test --manifest-path src-tauri/Cargo.toml
# Type check
pnpm build- Cross-platform support (Windows & Linux)
- Multi-language UI (i18n: Chinese + English)
- Theme switching (Light / Dark mode)
- OpenClaw health check & auto-repair
- Browser automation beyond Chrome (Firefox, Edge, etc.)
- Comprehensive CI/CD pipeline (multi-platform matrix)
- PR-level CI checks (lint, test, clippy)
- Architecture / sequence diagram
- Safer port management (avoid killing unrelated processes)
- Operator WebSocket graceful shutdown on disconnect
- Unified config source of truth (frontend vs backend)
OpenClaw Connector is built on excellent open-source projects:
- OpenClaw — The AI agent runtime
- Tauri — Lightweight desktop framework
- React — UI component library
- Zustand — Lightweight state management
- Tailwind CSS — Utility-first CSS framework
- Lucide — Beautiful icon set
See CONTRIBUTING.md for guidelines.
See SECURITY.md for the security model and vulnerability reporting.
