OpenClaw turns a self-hosted Gateway into the control center for chat-driven and scheduled automation. Install it on a computer or server, connect an AI provider, run the Gateway as a service, pair a channel such as Telegram or Discord, and then combine schedules, webhooks, skills, plugins, browser tools, and device nodes into controlled workflows. This guide covers the complete setup, always-on hosting choices, security boundaries, and troubleshooting.
Antihost offers VPS plans with one-click OpenClaw, NVMe storage and a 14-day free trial.
What OpenClaw is and how the Gateway works
OpenClaw is a self-hosted Gateway that connects chat applications to AI coding agents. Its architecture keeps sessions, routing rules, and channel connections in one runtime. The official documentation describes the Gateway as “the single source of truth for sessions, routing, and channel connections.”
As an Amazon Associate I earn from qualifying purchases.
You can run that Gateway on a personal computer, a Windows Hub, Linux, WSL2, a VPS or cloud instance, or an optional Raspberry Pi host. Once it is running, the local Control UI is available at http://127.0.0.1:18789/. A chat message, webhook, or timer can start a session; routing selects an agent and workspace; permissions determine which tools are available; and the result can return to the originating channel.
For a video-game workflow, that might mean a Discord request to inspect a server log, a nightly cron job that summarizes player activity, or a deployment webhook that asks an agent to verify a new build. OpenClaw supplies the plumbing, but the exact third-party action still depends on the channel plugin, provider, credentials, and tool permissions you configure.
#1 Best Overall
- Includes Raspberry Pi 5 with 2.4Ghz 64-bit quad-core CPU (8GB RAM)
- Includes 128GB Micro SD Card pre-loaded with 64-bit Raspberry Pi OS, USB MicroSD Card Reader
- CanaKit Turbine Black Case for the Raspberry Pi 5
- CanaKit Low Noise Bearing System Fan
- Mega Heat Sink - Black Anodized
Prerequisites and decisions to make first
- Node.js: use Node.js 24.16 or newer, or Node.js 26.1 or newer. Node 26 is the recommended branch. Recheck compatibility after upgrades because supported versions and defaults can change.
- AI access: have an existing Claude Code or Codex CLI login, or an API key for a supported provider. A detected login can be reused during Quick start; otherwise choose manual provider setup.
- A host: choose a desktop for experimentation, a dedicated Linux or Windows machine for local always-on operation, WSL2 when Windows is your main environment, or a VPS/cloud instance when remote uptime matters.
- A channel credential: Telegram requires a bot token and is documented as the quickest first channel. Other channels require their own app, bot, account, or plugin credentials.
- A security boundary: decide whether the Gateway will serve only you or other people. Main-session tools run on the host unless sandboxing is configured, so do not expose a new installation to untrusted senders without pairing and tool restrictions.
The getting-started documentation says initial setup takes about five minutes. Treat that as a documentation estimate, not a guaranteed completion time; provider authentication, channel approval, and host hardening can take longer.
Install OpenClaw
macOS or Linux
Run the documented installer from a shell:
curl -fsSL https://openclaw.ai/install.sh | bashAfter installation, open a new shell if your package manager changed the PATH. Confirm that the openclaw command is available before continuing.
Windows PowerShell
In PowerShell, use:
iwr -useb https://openclaw.ai/install.ps1 | iexUse a normal user account where possible. If Windows policy blocks the installer, use the documented Windows Hub or WSL2 route rather than weakening system policy without understanding the consequence.
Other installation paths
OpenClaw also documents npx openclaw@latest, npm, pnpm, bun, Docker, Nix, desktop companion applications, WSL2, and source installation. The source-development path is:
git clone <repository>
pnpm install
pnpm build
pnpm ui:buildThe repository specifically states that plain npm install at its root is not supported. Use the package-manager path documented for the release you are installing instead of mixing commands from different installation methods.
Complete onboarding and provider authentication
Start onboarding after the CLI is installed. The wizard offers two useful modes:
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Quick start
Choose Quick start when OpenClaw detects a usable Claude Code or Codex CLI login, or another supported access path. It reuses the detected credential and verifies it with a real completion. A successful login check proves that the provider path works; it does not grant channel permissions or host-tool access.
Rank #2
- Includes Raspberry Pi 5 16GB with 2.4Ghz 64-bit quad-core CPU (16GB RAM)
- Includes 128GB Micro SD Card pre-loaded with 64-bit Raspberry Pi OS, USB MicroSD Card Reader
- CanaKit Turbine Black Case for the Raspberry Pi 5
- CanaKit Low Noise Bearing System Fan
- Mega Heat Sink - Black Anodized
Custom setup
Choose Custom setup when you need to select a provider, enter an API key, choose a workspace, or set options deliberately. If the detected route fails, switch to manual provider setup rather than repeatedly retrying an invalid credential.
You can open the classic wizard explicitly with:
openclaw onboard --classicFor later targeted changes, use:
openclaw configureRecord which provider and account the Gateway uses. That makes later cost, quota, and incident checks easier, especially when several agents or channels share one host.
Install the Gateway as an always-on service
Foreground onboarding is useful for the first test, but it is not a durable deployment. Stop the foreground process with Ctrl+C, then install and inspect the service:
Free tools Windows power users keep installed
One-click scans. No signup required.
- Run
openclaw gateway install. - Run
openclaw gateway status. - Run
openclaw dashboard.
The service mapping depends on the operating system:
| Platform | Service mechanism | What to check |
|---|---|---|
| macOS | Per-user LaunchAgent | The Gateway starts for the configured user and listens on port 18789. |
| Linux or WSL2 | Per-user systemd unit | The user service is enabled and remains available after the shell closes. |
| Native Windows | Scheduled Task | The task is created for the intended account; a per-user Startup-folder fallback is used if task creation is denied. |
openclaw gateway status should report a Gateway listening on port 18789. Open the dashboard with openclaw dashboard, send a message in the Control UI, and confirm that an agent replies. This isolates provider and Gateway problems before you add a channel.
Connect a chat channel
Telegram: the simplest first channel
Telegram is documented as the fastest first channel because the setup centers on a bot token. Use the guided or scripted channel command:
openclaw channels addChoose Telegram in the prompts, enter the bot credential, and keep the token private. Send a test message only after the Gateway dashboard works locally.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Rank #3
- CanaKit Raspberry Pi 5 Essentials Starter Kit
Discord and other channels
The official channel list includes Discord, Slack, WhatsApp, Signal, iMessage, Google Chat, Microsoft Teams, Matrix, Mattermost, Zalo, and additional plugin channels. The same openclaw channels add command provides guided or scripted configuration, but each service has different credential, permission, media, and event-delivery requirements.
| Channel choice | Credential setup | Access control | Media and groups | Maintenance note |
|---|---|---|---|---|
| Telegram | Bot token | Pair unknown direct-message senders before allowing access | Suitable for text and supported bot media; group permissions depend on bot settings | Officially documented as a quick first channel |
| Discord | Application or bot credentials and server permissions | Restrict servers, channels, roles, and direct messages | Useful for game communities; verify attachment and thread behavior for the installed plugin | Permission changes in the server can alter what the bot sees |
| Slack, WhatsApp, Signal, iMessage, Google Chat, Microsoft Teams, Matrix, Mattermost and others | Service-specific account, app, or plugin credentials | Apply the platform’s membership and DM rules, then OpenClaw pairing where supported | Support varies by channel plugin; do not assume identical media or group features | Check whether the plugin is official or community-maintained |
Channel counts and capabilities change. The integrations directory lists 29 chat channels in 2026, but that is a page count rather than a promise that every channel has identical features or support status.
Pair senders before accepting automation requests
OpenClaw’s security guidance says to “Treat inbound messages as untrusted input.” DM-capable channels pair unknown senders by default. When a legitimate person receives a pairing request, approve it with:
openclaw pairing approve <channel> <code>Pairing answers who may start a session; it does not make every request safe. A paired sender can still ask the agent to access sensitive files or invoke a dangerous tool, so combine pairing with sandboxing, workspace restrictions, and an allowlist of tools.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteAssemble an end-to-end automation
OpenClaw groups its automation surfaces into cron (also called automations), hooks, webhooks, and transcripts. Skills load repeatable procedures on demand. Plugins add channel, provider, and tool surfaces. Browser tooling gives an agent a dedicated browser. Nodes expose capabilities such as camera, screen, voice, location, and Canvas. Multi-agent routing separates sessions by agent, workspace, or sender.
A robust workflow follows this order:
- Trigger: receive a chat message, webhook, or scheduled event.
- Route: select the agent, workspace, and session policy.
- Authorize: allow only the tools needed for this particular job.
- Execute: use a skill, plugin, browser, node, or host command to perform the work.
- Respond: send a concise result, error, or approval request back to the originating channel.
- Record: retain the transcript or operational log needed to diagnose the run.
Scheduled game-server report
A cron automation can run a nightly report for a private game server. The schedule starts an agent session; the agent reads only the designated log directory, applies a reporting skill, and posts a summary to a restricted channel. Do not give that reporting agent unrestricted shell access when read-only analysis is sufficient.
Webhook-driven build check
A build system can call a webhook after a game update is published. Route the event to a release-check agent, allow access to the deployment workspace and browser tool, and return a pass, failure, or request for human approval. Validate webhook signatures and reject unexpected payloads before they reach the agent.
Rank #4
- All-in-One Complete Kit: This SANOOV RPi 5 bundle comes with Raspberry Pi 5 4GB RAM single board, active cooler, durable ABS case and screwdriver. No extra parts needed, ready to use right out of the box for beginners and hobbyists
- Powerful Single Board Computer: Equipped with 4GB RAM and high-performance processor, delivers fast running speed for 4K playback, AI projects, programming and daily computing tasks. SANOOV for raspberry pi 5 4GB is equipped with broadcom 64 quad-core Arm Cortex A76 processor with gigabit ethernet and upgraded with IEEE 802.11ac Wi-Fi, Bluetooth 5.0 dual-band 2.4Ghz and 5Ghz and Power Over Ethernet (POE). Upgrading delivers 2-3 x speed vs Pi 4, redefining the experience
- Efficient Active Cooler: Effectively lowers operating temperature and prevents performance throttling. Runs quietly even under long-time heavy load, ensures stable operation all day long. SANOOV RPi 5 4GB kit offer an active cooler, which combines an aluminium heatsink with a high-performance PWM fan. Active cooler is fully compatible with the Pi OS, which can effectively reduce the temperature of RPi5 and ensure its good performance during long-term high load operation
- Sturdy ABS Protective Case: Well-fitted for Raspberry Pi 5 board, can be secured with 4 screws to effectively protect the Pi 5 motherboard from damage, reserves full access to all ports and buttons. SANOOV uses ABS material to produce the case, which has a softer texture and feel. Meanwhile, SANOOV case adopts a layered design for easy disassembly and installation. (Tip: The Case cannot install M.2 HAT Add on Board and Solid State Drive!)
- Wide Application & Full Compatibility: Seamlessly compatible with official OS and mainstream peripheral accessories for Raspberry Pi 5. Whether you are a beginner, student, electronics hobbyist or professional developer, this all-in-one kit meets your diverse needs. It excels in IoT projects, robotics design, retro gaming devices, home media servers and other DIY creations. Backed by a large global community, you can easily find guides, technical support and shared projects online
Chat-triggered support task
A moderator can ask in Discord or Telegram for the current server status. Route that channel to a support agent with a narrow status skill and read-only node or plugin access. If the request needs a restart, require an explicit approval step instead of allowing a normal chat message to execute it immediately.
Hooks, skills, and transcripts
Use hooks for lifecycle reactions, skills for reusable procedures, and transcripts for auditing and troubleshooting. Keep skills small and explicit: list their inputs, permitted tools, expected output, and failure behavior. A transcript should help you identify which trigger, route, and tool produced a result without exposing secrets in the channel.
Choose a host for continuous operation
| Deployment | Setup effort | Always-on reliability | Exposure and remote access | Headroom and maintenance | Best fit |
|---|---|---|---|---|---|
| Local desktop | Lowest for initial learning | Depends on sleep, reboots, and the user’s login session | Keep the Gateway local; use a deliberate remote-access method rather than opening the port directly | Usually good CPU, memory, and storage, but the machine is shared with other work | Testing, personal automations, and development |
| Windows Hub | Moderate; uses the Windows-specific service path | Good when configured as a dedicated, powered-on host | Apply Windows account and firewall controls | Maintenance follows Windows updates and scheduled-task behavior | Windows-first households or studios |
| Linux service | Moderate | Strong when the systemd user service and power policy are configured correctly | Expose only through a controlled network path; keep the Gateway port private | Flexible storage and memory; requires Linux administration | Dedicated local automation host |
| WSL2 | Moderate on a Windows workstation | Depends on the Windows host staying awake and WSL2 remaining available | Separate Windows and Linux networking considerations apply | Convenient Linux tooling, but less independent than a dedicated server | Developers who need Linux tools on Windows |
| VPS or cloud | Higher because you must harden and maintain a remote system | Usually best for unattended operation, subject to the provider and your monitoring | Internet exposure is the main risk; use firewall rules, private access, and pairing | Choose enough memory and storage for agents, transcripts, browser sessions, and plugins; patch the OS | Remote teams and 24/7 workflows |
| Raspberry Pi host | Moderate hardware and Linux setup | Can be power-efficient for a small Gateway, but depends on storage, cooling, and workload | Keep it on a private network or controlled tunnel | Less CPU, memory, and storage headroom than a desktop or VPS; no official preferred model or benchmark is specified | Small, local, always-on installations |
The documentation names Fly.io, Hetzner, GCP, Azure, exe.dev, and EasyRunner among VPS or cloud options. Selecting one is an infrastructure decision, not an OpenClaw performance guarantee; compare region, backups, firewall controls, memory, storage, and operating cost separately.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Secure an exposed or shared Gateway
- Keep the control surface private: do not publish port 18789 directly to the internet unless you have a specific, hardened access design.
- Pair unknown senders: approve only expected users and channels.
- Sandbox tools: main-session tools run on the host by default. Configure sandboxing before allowing other people or public channels to invoke host tools.
- Separate workspaces: route personal, game-server, and administrative tasks to different agents or workspaces where possible.
- Minimize credentials: give each plugin only the token or account scope it needs, and never paste secrets into a chat prompt.
- Validate webhooks: authenticate the sender and reject malformed or replayed events.
- Back up configuration: preserve the files and credentials needed to restore the Gateway, but store backups outside publicly readable locations.
Operate and update OpenClaw
After the first successful message, learn the operational command families rather than waiting for a failure:
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
openclaw doctorchecks common configuration and environment problems.openclaw statusreports overall service state.openclaw healthchecks health information.openclaw logshelps trace service and automation failures.openclaw backupsupports configuration preservation.openclaw migrateassists with documented migrations.openclaw updateapplies updates through the supported path.
Before updating, record the current Node version, provider route, channel plugins, and custom skills. Recheck Node compatibility and release notes after every update because versions, defaults, and integration behavior are volatile.
Troubleshoot the common failure points
The command is not found after installation
Open a new terminal, verify the package’s bin directory is on PATH, and confirm that you did not mix a source checkout with a global installation. If you used the source path, use its documented pnpm build steps rather than running root-level npm install.
Onboarding cannot complete a provider request
Check the provider login or API key independently, then rerun openclaw configure. A channel token cannot fix a provider-authentication failure; solve the AI access path first.
Best Value
- 【What you Get】You will get 1*Pi 5 8GB Single Board,1*RasTech Case,1*Active Cooler,1*Screwdriver,1*Installation instructions,12-month free warranty, lifetime service, 24-hour prompt and friendly response.
- 【More Connectors】There are two USB 3.0 ports(5Gbps simultaneously) and two USB 2.0 ports, which triple total bandwidth ,support any combination of up to two cameras or displays. Peak SD card performance is doubled through support for the SDR104 high-speed mode. It provides a smooth desktop experience for you. Offer Gigabit Ethernet and a PCIe interface, along with dual-band Wi-Fi and Bluetooth 5.0/BLE wireless capability. The RasTech Pi 5 Kit use the new 27W 5.1V 5A USB-C power connector.
- 【 Support Dual 4Kp60 Display 】Each of the two microHDMI sockets can control a 4K display at 60 Hertz, now support HDR, offering super HD video for media streaming projects. RPi 5 is the first RPi model that comes with a PCI Express port (PCIe 2.0 x1 with 500 MB/s) to attach SSDs (requires separate M.2 HAT).
- 【 Excellent Chips And Applications】Pi 5 is a full-size Pi computer using silicon built in-house at Pi. The RP1 “southbridge” provides the bulk of the I/O capabilities for Pi 5. Pi 5 is more friendly and convenient in the development of Internet of Things, Web development, machine identification, automatic control and other electronic equipment applications and network.
- 【 Faster CPU, Better GPU 】 Pi 5 features a Broadcom BCM2712 64-bit quad-core Arm Cortex-A76 processor running at 2.4GHz, it delivers a 2–3× increase in CPU performance relative to RaspberryPi 4. The 800MHz VideoCore VII GPU is compatible to OpenGL ES 3.1 and Vulkan 1.2, substantial uplift in graphics performance. Pi 5 Offers lightning-fast CPU speed, a PCI Express interface, a Real Time Clock (RTC) and a power button and runs significantly cooler than Pi 4.
The Gateway service is not listening
Run openclaw gateway status, inspect openclaw logs, and confirm that port 18789 is not blocked by another process or firewall rule. On Linux and WSL2, check the per-user systemd unit. On Windows, inspect the Scheduled Task or the Startup-folder fallback.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minutePC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11The dashboard works but the channel is silent
Verify the channel credential, plugin status, bot permissions, and pairing state. Send a direct test message before testing a group. In a game-community server, confirm that the bot can read the selected channel and that role or thread restrictions have not changed.
An automation performs too much
Stop the workflow, revoke unnecessary credentials, and narrow its route and tool allowlist. Move risky actions behind a human approval step and configure sandboxing before enabling the automation again.
How large is the OpenClaw integration ecosystem?
The official integrations directory currently lists 29 chat channels, 64 model and media providers, and 142 official plugins, all labeled as OpenClaw 2026 page counts. These numbers can change. The same directory lists macOS, Windows, iOS, and Android companion nodes and community skills or plugins for services including GitHub, Notion, Obsidian, Home Assistant, Philips Hue, Spotify, Sonos, email, X/Twitter, 1Password, and weather.
“Official” and “community-maintained” are different risk categories. Before adding a plugin, check its maintenance status, requested permissions, credential handling, and compatibility with your installed OpenClaw release.
Recommended Free Tools
A practical launch checklist
- Install Node.js 24.16+ or 26.1+; prefer Node 26.
- Install OpenClaw using the macOS/Linux, Windows, package-manager, container, or source path appropriate to your host.
- Run onboarding and verify a real provider completion.
- Stop the foreground process and run
openclaw gateway install. - Confirm
openclaw gateway statusreports port 18789. - Open the dashboard and send a local test message.
- Add Telegram or another channel with
openclaw channels add. - Pair expected senders and deny unknown requests.
- Build one narrow automation before adding schedules, webhooks, browser actions, nodes, or multiple agents.
- Apply sandboxing, least-privilege tools, backups, logs, and an update plan before sharing the Gateway.
That sequence gives you a working conversational path first, then adds automation one controlled surface at a time. It is safer and easier to diagnose than exposing every integration during the initial install.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
