--- title: "Deploy Claw3D" description: "A 3D virtual office for OpenClaw, Hermes, and custom AI agent runtimes." category: "AI/ML" url: https://railway.com/deploy/claw3d --- # Deploy Claw3D A 3D virtual office for OpenClaw, Hermes, and custom AI agent runtimes. **[Deploy Claw3D on Railway](https://railway.com/template/claw3d)** - **Creator:** INF Labs - **Category:** AI/ML - **Total deploys:** 1 ## Template content ### claw3d https://hebbkx1anhila5yf.public.blob.vercel-storage.com/logo-iBhp9t3RItzx9SoNobfuPQChIm022x.jpg - **Image:** ghcr.io/iamlukethedev/claw3d:main - **Start command:** `/bin/sh -c 'npm install --no-save typescript@5 && exec node server/index.js'` - **Public domain:** Yes ## Documentation # Deploy and Host Claw3D on Railway Claw3D is a live 3D workspace for AI agents. Instead of watching agent activity through logs and dashboards, Claw3D lets you step into a virtual office where agents collaborate, review work, run tasks, and interact in real time. It can connect to OpenClaw, Hermes, demo runtimes, and custom agent backends. ## About Hosting Claw3D Hosting Claw3D on Railway gives you a self-hosted visual workspace for AI agent systems. This template runs the official Claw3D container and exposes the Studio interface through Railway. Claw3D acts as the visualization and interaction layer, while the actual agent runtime remains in the connected backend. The Studio can connect to OpenClaw Gateway, Hermes, a built-in demo runtime, or a custom HTTP runtime. Gateway credentials and runtime settings can be configured after deployment, so Claw3D can be launched even before a real agent backend is ready. ## Common Use Cases * Watch AI agents work together in a live 3D office * Visualize multi-agent collaboration and task execution * Monitor agent activity, sessions, approvals, and workflows * Review development and operational activity from one workspace * Connect Claw3D to OpenClaw or Hermes * Explore the interface using demo mode * Use custom agent runtimes through the Claw3D runtime layer ## Why Use Claw3D? | Capability | Claw3D | | ----------------------------------------- | :----: | | Live 3D virtual office | ✅ | | Multi-agent visualization | ✅ | | OpenClaw support | ✅ | | Hermes support | ✅ | | Custom runtime support | ✅ | | Built-in demo mode | ✅ | | Agent chat and interaction | ✅ | | Runtime session monitoring | ✅ | | Workflow visualization | ✅ | | GitHub and development-oriented workflows | ✅ | | Custom office layouts | ✅ | | Browser-based Studio | ✅ | | Separate database required | ❌ | | Redis required | ❌ | | External object storage required | ❌ | Claw3D is designed for users who want to understand agent activity visually instead of relying only on logs, terminal output, or traditional dashboards. ## Dependencies for Claw3D Hosting * Official Claw3D container image * Optional OpenClaw Gateway * Optional Hermes runtime * Optional custom agent runtime * A persistent filesystem if you want Studio settings to survive redeployments ### Implementation Details Claw3D uses a Studio-first architecture. ```text User Browser │ │ HTTP / WebSocket ▼ Claw3D Studio :3000 │ │ Server-side Proxy ▼ Agent Runtime │ ├── OpenClaw ├── Hermes ├── Demo Gateway └── Custom Runtime ``` The browser communicates with Claw3D Studio. Claw3D Studio then establishes the upstream connection to the selected agent runtime. This keeps the runtime connection logic inside the Studio layer instead of requiring the browser to connect directly to the agent backend. ## Supported Runtime Backends Claw3D supports multiple runtime modes: | Runtime | Supported | | ------------------------- | :-------: | | OpenClaw Gateway | ✅ | | Hermes | ✅ | | Built-in Demo Gateway | ✅ | | Custom HTTP Runtime | ✅ | | Local runtime | ✅ | | Claw3D-compatible runtime | ✅ | This makes Claw3D useful even if you are not using OpenClaw exclusively. ## Connecting to OpenClaw Claw3D can connect to an OpenClaw Gateway using a WebSocket endpoint. For example: ```text ws://gateway-host:18789 ``` or when the Gateway is exposed securely: ```text wss://gateway.example.com ``` The Gateway URL and authentication token can be configured after Claw3D is deployed. If the Gateway runs inside the same Railway project, Railway private networking can be used for the connection. If the Gateway runs externally, use a WebSocket endpoint that is reachable from the Claw3D service. ## Custom Runtime Support Claw3D can also connect to custom orchestrators and agent runtimes through its HTTP runtime interface. A typical custom runtime provides endpoints such as: ```text GET /health GET /state GET /registry POST /v1/chat/completions ``` Claw3D proxies these requests through its own runtime layer, avoiding direct browser-to-runtime communication. This allows Claw3D to act as a visual frontend for custom multi-agent systems beyond OpenClaw and Hermes. ## Public Access Claw3D Studio listens on: ```text 3000 ``` Generate a Railway public domain for port `3000`. Example: ```text https://your-claw3d.up.railway.app ``` Open this URL to access the Claw3D Studio and configure your runtime connection. ## Studio Security Claw3D supports a Studio access token when the interface is exposed publicly. This is strongly recommended for Railway deployments because the Studio provides access to agent connections, runtime configuration, and operational information. The access token protects the Studio itself before a user can interact with the connected AI runtime. ## Persistence Claw3D stores Studio configuration and preferences on the local filesystem. Persistent data can include: * Gateway connection settings * Selected backend mode * Agent preferences * Focused-agent state * Desk assignments * Office layout state * Other Studio preferences A persistent Railway volume is recommended if you want these settings to survive redeployments. The default Claw3D Studio configuration path is based under: ```text ~/.openclaw/claw3d/ ``` ## Database and Cache Requirements Claw3D does not require a traditional database or cache for its default deployment. | Dependency | Required | | ----------------------- | :---------: | | PostgreSQL | ❌ | | MySQL | ❌ | | MongoDB | ❌ | | Redis | ❌ | | External Object Storage | ❌ | | External Cache | ❌ | | Persistent Filesystem | Recommended | This keeps the deployment lightweight while still supporting multiple agent runtimes and a full 3D interface. ## Claw3D vs Traditional Agent Dashboards | Feature | Claw3D | Traditional Dashboard | | -------------------------------- | :----: | :-------------------: | | Live 3D agent workspace | ✅ | ❌ | | Visual agent movement | ✅ | ❌ | | Multi-agent collaboration view | ✅ | Limited | | Runtime monitoring | ✅ | ✅ | | Agent chat | ✅ | Depends | | OpenClaw integration | ✅ | Depends | | Hermes integration | ✅ | Depends | | Custom runtime support | ✅ | Depends | | Office layout customization | ✅ | ❌ | | Demo mode | ✅ | Rare | | Immersive workflow visualization | ✅ | ❌ | The main difference is that Claw3D treats the agent system as a workspace rather than a traditional monitoring screen. ## Getting Started After Deployment 1. Deploy the Claw3D template. 2. Wait for the Studio service to become available. 3. Generate a Railway public domain for port `3000`. 4. Open the Claw3D Studio URL. 5. Authenticate with the Studio access token. 6. Select your runtime type. 7. Configure your OpenClaw, Hermes, demo, or custom runtime connection. 8. Connect to the runtime. 9. Open the 3D Office and start monitoring your AI agents. If you do not have an agent runtime yet, use the built-in demo mode to explore the Claw3D interface first. ## Why Deploy Claw3D on Railway? Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it. By deploying Claw3D on Railway, you can turn your AI agent infrastructure into a live, interactive 3D workspace while keeping the Studio, networking, and runtime connections within a single deployment platform. ## Similar templates - [Chat Chat](https://railway.com/deploy/-WWW5r) — Chat Chat, your own unified chat and search to AI platform. - [stella](https://railway.com/deploy/stella) — Self-host stella with web, API, Postgres, Redis, and object storage. - [Hermes Agent | OpenClaw Alternative with Dashboard](https://railway.com/deploy/hermes-agent-or-openclaw-alternative-wit) — Self-Hosted Hermes AI Agent for Telegram, Discord & Slack Open this page in a browser: https://railway.com/deploy/claw3d